Imagine growing up in a suburban, working and middle-class town. In elementary school, you sit through all the anti-drug assemblies with your peers while you roll your eyes because “no one I know does that.” Fast forward to a couple years after your high school graduation. You start getting phone calls about old friends dying. It seems like you’re getting a call every other week. They all died of the same thing- overdose. Some developed an opioid habit after an injury, while others shot up once just to see what would happen and it stopped their heart. This is becoming the norm for communities across America, and it seems the epidemic is creeping into California.
California currently has the 7th lowest drug overdose death rate in the U.S., however the opioid epidemic has already devastated areas like Eureka and other towns in Humboldt County. Residents in these areas are citing lack of treatment options and limited affordable options as reasons for opioid abuse and misuse becoming out of control. Opioid use is leading to a rise in crime and a plethora of discarded needles waiting to be stepped on around town. Some residents have opened needle exchange shops and soup kitchens, but their good will isn’t enough.
What if we could predict where the opioid epidemic would travel to next? What if we could get ahead of it before addiction and overdose devastates another community? Prediction coupled with improved interventions could drastically decrease the opioid overdose death rate around the country.
In a paper recently accepted for publication in the journal, Drug and Alcohol Dependence, UCIPT affiliates found evidence that Google search term data can be used to predict future emergency department admissions for heroin use. They found that by looking at opioid-related search terms, they can predict the next years heroin-related emergency department visits for that region. By tracking opioid-related search terms, we can use this tool to identify hotspots for opioid use and intervene before it gets out of control. In the future, other forms of social data can be used as well like Twitter or Facebook.
What do you think of tools like this? How do you feel about the ethics of mining search term data?