Binge drinking in teens and young adults is a serious health concern, but there could soon be an app to help tackle the problem.
Researchers are developing a new intervention strategy that uses smartphones to curb excessive drinking.
They presented their findings in Colorado this month at the 40th annual scientific meeting of the Research Society on Alcoholism.
“Imagine a software program that can recognize when you were out drinking and send you customized support messages to remind you of your goals or offer safer transportation alternatives,” Dr. Brian P. Suffoletto, lead researcher, and assistant professor of emergency medicine at the University of Pittsburgh, told Healthline.
Suffoletto and his team recruited 38 young adults and monitored data from their cell phones.
Then, using computer learning, they built a predictive model of when drinking would take place.
From there, they programmed an app to intervene electronically.
According to Suffoletto, the software can detect sessions of young adult drinking with 95 percent accuracy.
How the app works
So, how can a phone predict when someone is drinking?
Researchers compiled 20 features that could be identified with phone sensors in order to create a predictive model.
The model uses indicators such as time, day of the week — predictably evenings and weekends — that point to a much higher likelihood of drinking.
The software also uses some innovative indicators as well, such as the number and length of outgoing calls, text message errors and deletions, and even the number of times emoticons are used in text messages.
“Top phone sensor features included behaviors like keypress speed [slower during drinking] and errors [greater during drinking], duration of outgoing calls [longer during drinking], and change in activities [greater during drinking],” Suffoletto said.
The goal is to help predict when drinking will happen and then help users make better and more informed decisions.
Warning the user
Suffoletto explained that the app could respond to users' drinking in a variety of ways.
It could send a text message or push notification, cautioning the user.
Or it might remind them not to drive, or to use a taxi or ride-sharing app like Uber.
There are any number of strategies that could be deployed, perhaps even customized by the user.
“With the ability to detect when an individual is drinking, we will then be able to test any number of interventions delivered at those times, which could include blocking certain phone features [e.g., Tinder, Facebook], delivering supportive messaging, or contacting supportive friends, family, or a counselor,” Suffoletto said.
Research, including some from the Centers for Disease Control and Prevention (CDC), indicates that binge drinking peaks in teens and young adults between the ages of 18 and 34. It is considered a serious but preventable public health concern.
New reports have shown that teen drinking, specifically among high school students, has been declining, but it is still problematic.
Using technology to combat drunk driving
As internet usage has spread, and smartphone technology has advanced in recent years, there has been a rise in so-called “e-intervention” or “digital intervention” strategies to communicate with drinkers and help them make better decisions.
Results have been mixed.
Portable smartphone breathalyzer products aimed at consumers have faced waves of criticism about their accuracy.
Recently, the Federal Trade Commission (FTC) entered into a lawsuit with one major manufacturer, Breathometer, about the company’s claims that it produced a “law-enforcement grade product.”
The FTC won that lawsuit.
Meanwhile, studies on ride-sharing apps like Uber and Lyft, sometimes hailed as the solution to drunk driving by giving drinkers easy access to transportation, offer differing conclusions about their effectiveness.
However, Suffoletto’s app is different than any current iterations of e-interventions.
“To our knowledge, this is the first time that phone sensors have been used to predict drinking occasions,” he said in a press release.
Through computer learning and data models, the app can help intervene before users are already too drunk to make good decisions. A reminder to get a taxi at 2 a.m. is helpful only to a certain extent. It doesn’t actually prevent excessive drinking.
By intervening before or during sessions of drinking, this software could cut closer to the root of the problem, rather than solely taking on its symptoms like drunk driving.
“[The] overall goal is to create efficient and effective interventions that will actually be used by young adults to reduce the preventable public health problems related to excessive alcohol consumption,” said Suffoletto.