9:15 - 10:45 amSunday, September 18
Lower Lobby
Machine Learning to Detect Prescription Opioid Abuse Promotion by Illicit Online Pharmacies
Lower Lobby
Machine Learning to Detect Prescription Opioid Abuse Promotion by Illicit Online Pharmacies
Assistant Professor/UC San Diego - School of Medicine
Introduction:  Non-medical use of prescription medications (NUPM) is a public health emergency. The national epidemic of prescription drug abuse is exacerbated by the presence of illicit online pharmacies... Read more

Description

Introduction:  Non-medical use of prescription medications (NUPM) is a public health emergency. The national epidemic of prescription drug abuse is exacerbated by the presence of illicit online pharmacies that illegally sell controlled substances directly to the consumer without requiring prescription and outside the oversight of a healthcare professional. Despite a Federal law that explicitly prohibits this practice, prior studies have uncovered promotion of illicit online pharmacy sales of controlled substance via the popular microblogging site Twitter.

Methods: We conducted surveillance of Twitter and collected tweets filtered for keywords associated with commonly abused prescription drugs over a five month period in 2014. We used a three step process that used unsupervised machine learning applied to the subset of tweets pertaining to each drug keyword in order to isolate and characterize tweets which promote illicit online pharmacies. In the first step, machine learning was applied to the corpus of tweets pertaining to a particular drug, and a set of topics were produced. In the second step, we annotated the topics for their relevance to illicit online pharmacy promotion (and irrelevant topics and associated tweets were discarded from further analysis). Within the topics marked as relevant, we further pruned the tweets in this topic to remove irrelevant tweets. Finally, we manually annotated characteristics of each identified tweet and any hyperlinks embedded in the tweet to determine and verify its classification in the promotion of online sale of controlled substances.

Results:  Using this methodology, we examined 619,937 tweets containing the keywords codeine, percocet, fentanyl, vicodin, oxycontin, oxycodone, and hydrocodone. A total of 1778 tweets were annotated as containing information relevant to illicit online drug sales (the % of tweets relevant for each controlled substance ranged between 40% - 95%) using our unassisted machine learning protocol. We detected a total of 1608 relevant tweets with imbedded hyperlinks, but only 46 (7 distinct URLs) were still “live” at the time of manual annotation. These hyperlinks directly linked to 2 “rogue” online pharmacies, 1 online pharmacy that could not be verified but sold drugs without a prescription, two marketing affiliates providing links to rogue online pharmacies, a purported wholesale drug company/distributor, and an online classified ad advertising the local sale of controlled substances illegally.

Conclusions:  This study presents an innovative methodology to filter large volumes of Twitter content and isolate tweets specifically promoting illegal online sale of controlled substances. Results from this study indicate that controlled substances are sold using various marketing techniques including tweets directly linking to an illicit online pharmacy, use of marketing affiliates, wholesale distribution, and even the use of local online classified ads for direct purchase from an individual seller. This methodology can be useful in detecting illicit online access points that can be reported to the U.S. Drug Enforcement Agency and U.S. Food and Drug Administration for immediate enforcement action as codified in the 2008 Ryan Haight Online Pharmacy Consumer Protection Act.

Timothy Ken Mackey is the Director of the Global Health Policy Institute (www.ghpolicy.org) and an Assistant Professor of Anesthesiology and Global Public Health at University of California, San Diego School of Medicine. He is also the Associate Director for the UC San Diego MAS Program in Health Policy & Law. He earned his BA in Political Science-International Relations from the UC San Diego, a Masters Degree from UC San Diego MAS Program in Health Policy & Law, and his PhD in the Joint Global Public Health program with UC San Diego and San Diego State University. He has also completed an Executive Course in Global Health Diplomacy at the Graduate Institute, Geneva.  His work focuses on a broad array of multidisciplinary topics in domestic and global public health research. This includes cross-cutting research in disciplines of public health, medicine, international relations, public policy, law, technology, economics and intellectual property, technology environmental health, eHealth, crime, and global governance. He has extensive professional working experience in the private sector and has worked for the World Health Organization and the US Department of State among others.

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search