How scientists are using Google searches to predict heroin overdose
We all know Google search helps with looking up websites, but, in a first, scientists are using the tech for tracking heroin overdoses even before they occur. It may sound a bit weird but they have actually created a model, which uses Google search data to predict how many cases are likely to occur, Quartz reported. Here's all about it.
How this model works?
The model uses Google searches for commonly used opioid-related terms, including slangs, to predict cases of overdoses from the drug. It is not being used officially, but in a test, the scientists found that their model could be used to explain 72% variation in year-after-year heroin-related hospital visits. They also found that the higher number of searches per keyword was linked to more overdoses.
In the test, the team used data for eight years
To show the working of the model, the team used opioid-related Google searches from nine metro cities in the US and heroin overdose-related hospital visits in those cities from 2004 to 2011. They used keywords like Avinza, Brown Sugar, China White, Codeine, Kadian, and Methadone.
The tech is developing, but its potential is massive
Moving ahead, this model could prove effective in preparing for, even avoiding, heroin overdose cases, deaths. However, as of now, it is important to note that the tech is at a nascent stage and it would need multiple improvements before being ready for deployment. For instance, the system would have to learn to differentiate between terms which may have two meanings like Brown Sugar.
Opioid overdose cases in India
In India, over a million opioid overdose-related cases occur every year. Meanwhile, in the US, the situation is even worse, with as many as 100 victims dying from overdose daily.
