Johns Hopkins Engineering | Summer 2013 Johns Hopkins University Whiting School of Engineering at Johns Hopkins University

f


Ferreting Out Illicit Drug Data

In recent years, recreational drug use has become particularly dicey, as young people turn to surprising new sources for a fix. After swallowing, snorting, smoking, or injecting drugs that turned out to be mephendrone—known colloquially as bath salts—some users even have died. More recently, synthetic marijuana, known as Spice, K2, and Mojo, has been linked to a handful of teen deaths and more than 10,000 hospitalizations in the United States alone. Experts contend that as many as one novel unidentified drug a week enters the global marketplace.

Conventional data-collection surveys—which rely largely on honest self-reporting and sometimes take up to a year to conduct— are proving ineffective at providing health workers with a quick, clear picture of who is using what, and where. “It’s a major problem,” says Mark Dredze, an assistant research professor in Hopkins’ Department of Computer Science. “A person comes into the emergency room overdosing, and you have no idea what they are taking. They don’t know what they are taking.”

That’s what happened to 16-year-old Emily Bauer of Cypress, Texas, who bought “potpourri,” or synthetic marijuana, from a local gas station and smoked it at a party. She ended up on a breathing tube and nearly died. Chase Burnett, a 16-year-old from Georgia, wasn’t as fortunate. He drowned in the family hot tub after smoking “Mojo Diamond EXTREME.” “It’s been a complete and utter mess trying to figure out what is going on,” Dredze says.

Dredze and PhD researcher Michael J. Paul, MS ’12, who have pioneered the use of sophisticated computer algorithms to mine social media sites for infectious diseases such as the flu, are close to filling gaps in much needed knowledge.

The team has developed a computer algorithm that ferrets out valuable information about illicit drug use. Using sophisticated natural language processing and machine learning, it crawls public posts in discussion forums to find associations between words that represent new drugs, uptake methods (such as snorting, sniffing, or smoking), and other aspects such as health impacts.

The goal: to provide doctors, policymakers, drug enforcement officials, and even drug users themselves with valuable, real-time information about the dangers of new chemical concoctions.

Their recent conference paper analyzed more than 400,000 public posts on drugforum.com about 22 new drugs, focusing on five emerging drugs that included bath salts and salvia. Some findings: Bath salts were linked to dehydration, teeth grinding, sweating, hot flashes, anxiety and paranoia, elevated heart rates and blood pressure, chest pain, and itchy rashes. Salvia, on the other hand, was linked to intense hallucinatory effects—including weird thoughts, feelings of unreality, bizarre non- Euclidian dimensions/geometrics, and feelings of floating.

A federal law, passed quickly in July 2012, is trying to curb use of these dangerous substances, but drug enforcement officials report continued use.

Next up: The team is working with addiction research faculty in the Department of Psychiatry and Behavioral Sciences at the School of Medicine to track new drugs in real time. Dredze says: “The pie-in-the-sky hope is that we will be able to identify a new drug within a month of its being introduced, identify what it is, who is taking it, and what the health hazards are.

— Mary Beth Regan

Page 1 of 1 pages for this article




Responses to "Ferreting Out Illicit Drug Data"

Leave a Reply:

* Name:

* Email:

* Comment:

* required

Remember my personal information
Notify me of follow-up comments?

Please enter the word you see in the image below:


JHU Engineering | Summer 2013