Young S, Zheng K, Chu L, Humphreys K. (2018). Internet searches for opioids predict future emergency department heroin admissions. Drug and Alcohol Dependence. (190): 166-169. doi: 10.1016/j.drugalcdep.2018.05.009
Researchers analyzed the relationship between volume of geolocated Google searches for prescription and non-prescription opioids (2004–2010) and annual heroin-related Emergency Department (ED) admissions (2005–2011) from The Substance Abuse and Mental Health Services Administration (SAMHSA) to determine whether opioid-related Google search data predicted next-year heroin-related ED admissions. Across 9 U.S. cities in the SAMHSA dataset, heroin-related ED admissions rose 12.8% from 1194.7 per 100,000 population (2005) to 1358.3 (2011). Since both prescription and non-prescription opioid use increase the risk of future heroin use, researchers entered the 12 prescription and non-prescription opioid-related keywords with the most Google searches (Opioid, Heroin, China White, Black Tar, Brown Sugar, Morphine, Fentanyl, Avinza, Codeine, Kadian, Methadone, and Oxymorphone) into Google Trends (a Google website that analyzes Google search queries across geographic regions) to calculate normalized search volume values for each keyword over the study period. Researchers modeled the relationship between the normalized volumes of opioid-related keyword searches and next-year heroin-related ED admissions. The best-fit model included 7 keywords (Avinza, Brown Sugar, China White, Codeine, Kadian, Methadone, and Oxymorphone) and revealed that volume of opioid-related keyword Google searches predicted 72% of variation in next-year heroin-related ED admissions. Keywords Avinza, Methadone, and China White exhibited the strongest associations between search frequency and next-year heroin-related ED admission. Results suggest that internet search data may be a valuable, low-cost tool in prediction of heroin-related ED admissions and may be of particular use in the many geographic regions with no current heroin use data.