For many years, researchers, healthcare professionals and government officials have relied upon the data collected from large, nationally representative self-report surveys as the “gold standard” for illicit drug use surveillance. Despite the ubiquitous use of self-report data for illicit drug-use surveillance, under-reporting is likely due to respondent self-presentation concerns, non-response rates (which can bias surveillance estimates) are increasing and most of these large, population-based surveys are conducted only once a year with results reported one to two years after the data were collected. Thus, the veracity of illicit drug use surveillance data is questionable, untimely, and is not sensitive enough to track new and emerging drug use trends.
Given the cost and potential ethical problems associated with more objective measures of drug use (i.e., collecting biological markers of illicit drug use), recent work has investigated using “big data” sources such as Twitter as a surveillance tool to track the epidemiology of illicit drug use in the United States and other countries. For example, in a recent study conducted by Hanson and colleagues (2013), the authors used Twitter to examine the tweets containing the keyword “Adderall” of individuals living near colleges and universities in several locations in the U.S. Termed “infoveillance,” (Eysenbach, 2009) the leveraging of “big data” social media platforms such as Twitter for the surveillance of illicit drug use (Chary et al., 2013; Cameron et al., 2013) may be the panacea to the problems that plague self-report survey data collection.
In response to a Request for Applications (RFA) from the National Institute on Drug Abuse to study drug use epidemiology using social media, Dr. Susan Woodruff and Dr. Mark Reed from the School of Social Work collaborated with an interdisciplinary group of colleagues at SDSU in Geography (Dr. Ming-Hsiang Tsou), the School of Communications (Dr. Meghan Moran and Dr. Brian Spitzberg), and Linguistics (Dr. Marc Gawron) as well as with Dr. Rouming Jin at Kent State University to draft a research proposal addressing the potential to leverage social media for illicit drug surveillance. This grant was submitted under the newly formed Human Dynamics in The Mobile Age research cluster at SDSU.
With Drs. Woodruff and Tsou as principal investigators, the team wrote a proposal requesting nearly two million dollars in funding for three years to develop and test a prototype system (Spatial, Temporal, and Regional Observation Network Generator for Drug Abuse Trend Analysis or STRONG-DATA system) designed to mine Twitter data for keywords related to several illicit (marijuana, cocaine, heroin, etc.) as well as prescription (Adderall, Oxycontin) drugs. This prototype system will incorporate machine learning processes, Geographic Information Systems (GIS), social network analysis, as well as geo-targeted social media data mining tools to assess both geographic and temporal trends of the use of several classes of illicit and prescription drugs. Additionally, the project proposes to validate the STRONG-DATA prototype with hospital emergency department data as well as data collected from a group of participant volunteers. Lastly, the proposal seeks to develop a secure and privacy-protected online dashboard for the real-time surveillance of drug use using the STRONG-DATA system. Translational research will also be conducted with healthcare providers and government officials to determine the utility and potential for sustainability of the online dashboard system. If funded, this project has the potential to improve current illicit drug use surveillance methods as well as enhance how communities and healthcare agencies respond and prepare for drug use epidemics (i.e., heroin overdoses in a community) or new and emerging drugs.
References
Chary, M., Genes, N., McKenzie, A., & Manini, A.F. (2013). Leveraging social networks for toxicovigilance. Journal of Medical Toxicology, 9(2), 184-191.
Cameron, D., Smith, G. A., Daniulaityte, R., Sheth, A. P., Dave, D., et al. (2013).PREDOSE: A semantic web platform for drug abuse epidemiology using social media. Journal of Biomedical Informatics, 46, 985-997.
Eysenbach, G. (2009). Infodemiology and infoveillance: Framework for an emerging set of public health informatics methods to analyze search, communication and publication behavior on the Internet. Journal of Medical Internet Research, 11(1), e11.
Hanson, C. L., Burton, S. H., Giraud-Carrier, C., West, J. H., Barnes, M. D., & Hansen,B. (2013). Tweaking and tweeting: Exploring Twitter for nonmedical use of apsychostimulant drug (Adderall) among college students. Journal of Internet Medical Research, 15(4), e62, 1-12.
Dr. Mark B. Reed is an Associate Professor in the SDSU School of Social Work and is a Research Fellow in the Center for Alcohol and other Drug Studies.