Data mining to predict addiction in teens

How can we use technology to help predict (and prevent) addiction? An interesting perspective with ideas on how to get ahead of substance abuse, before it even happens.

minute read

We Want to Protect Our Kids

Every parent fears that their teen will get hooked on drugs. This fear isn’t unfounded. Data from the National Institute on Drug Abuse (NIDA) shows that:

  • 12% of 8th graders
  • 26% of 10th graders
  • 38% of 12th graders

…say they have used an illicit drug at least once in the past year. While these numbers show a decline from previous years, there’s still cause for concern considering teens who abuse drugs are not only likely to perform poorly academically or drop out of school altogether but are also at a higher risk of getting unplanned pregnancies, infectious diseases (especially STIs) and becoming violent.

While your teen might show some tell-tale signs of drug use, pinning them down about this destructive habit can be difficult given teens’ natural reticence. However, this may soon change as researchers are looking into ways of mining information from social media to help predict teenage substance abuse.

Social Media Use Among Teens

There is no doubt that teen’s use of social media has blown up in recent years, largely facilitated by mobile devices. According to the Pew Research Center, 71% of teens use more than one social media site whenever they go online, Facebook being the top choice followed by Snapchat, Instagram and Twitter. This means that your teen might be quite chatty online even if they’re uncommunicative toward you.

While social media can be quite addictive on its own, studying what teens are sharing, liking and talking about can give fascinating insights into their lives and even reveal information that might point to substance abuse.

Mining For Substance Abuse

There is an abundance of information hidden within social media chatter. Researchers, using machine learning and text mining algorithms, are studying it to come up with automated prediction systems to uncover drug abuse trends among teens. Data from various social media sites is:

  1. Collected.
  2. Filtered.
  3. Mapped and aggregated .
  4. Then, analyzed to extract relevant information.

The premise behind this is that teens who use drugs have similar patterns of behavior as well as similar social network structures. Studies have shown that they tend to use certain words when posting online, like or follow particular social media pages or accounts and even share their experience of drug usage with each other.

Soon We’ll Be Able To Use Data Mining In Social Media To Predict Teenage Substance Abuse

This information, gathered via data mining, can be an invaluable resource in predicting:

  • The level of drug use
  • Types of drugs used
  • Side effects of drug use
  • Risk factors of drug use

…as well as behavior traits and patterns associated with drug use among teens. The information can also enhance the screening, identification, prevention and treatment of substance abuse and addiction in teenagers.

What About Privacy?

It’s impossible to talk about data mining without addressing the privacy and security concerns raised by such large scale aggregation of personal information. Once analyzed, the vast data shared by teens (and other social media users) can reveal a lot about an individual’s personal life, habits, likes and preferences. As such, data mining can be extremely intrusive.

Additionally, teens sometimes may not have given consent for their information to be used in data mining and may be unaware it’s being analyzed to predict substance abuse. This raises the question:

Should social media sites should be allowed to share users’ information without their knowledge and consent to begin with?

Another privacy issue with data mining is using the information gathered to stereotype or profile teens based on age, ethnicity, gender or even substance abuse. Although the data used is usually anonymized, it can still be linked to a large sub-population of teens e.g. those of a certain gender or race or those who live in a specific geographic region. Such conclusions can therefore lead to the wrongful discrimination or exclusion of certain groups of teens by schools, employers, insurance providers etc. should they get access to the mined information.

Getting Educated Yourself

While social media can be an incredible communication tool, it does have a dangerous dark side. It can perpetuate drug use and abuse. It can even celebrate it!

As a parent, you should strive to monitor your teens’ social media use. Further, you can educate them on the hazards posed by substance abuse. Where can you start? Here are some websites that can help you get the conversation started.

  1. NIDA search for “parents” + “teens”:
  2. NIDA for Parents:
  3. SAMHSA Parent Resources for Talking About Alcohol:
  4. Drug Parent Talk Kit:

Where Will Data Monitoring Take Us?

While the advantages of data mining are clear, at least in predicting teen substance abuse, it would be foolhardy to ignore the privacy issues that come along with it. There’s a need for clear policies and legislation that protect the privacy of social media users while availing relevant information for use in the interest of public health and safety. So, as we continue to try to prevent addiction, parents can play their part by monitoring their teens’ social media use and educating them on the hazards posed by substance abuse.

All in all, it will be interesting to see how all this plays out in future.

About the author
Tyler is a freelance writer/journalist, with past experience as the head content writer and outreach coordinator for HelpYourTeenNow. His areas of focus include: parenting, education, social media, addiction, and issues facing teenagers today.
I am ready to call
i Who Answers?