Google Trends Data: A Potential New Tool for Monitoring the Opioid Crisis

Abstract

<b><i>Introduction:</i></b> There is a need to strengthen the standard surveillance of the opioid overdose crisis in the USA. The role of Google Trends (GT) was explored in this context. <b><i>Methods:</i></b> In this study, a systemic GT search was done for a period from January 2004 to December 2018. “Naloxone” and “drug overdose” were chosen as search inputs. By using locally weighted scatterplot smoothing, we locally regressed and smoothed the relative search data generated by the GT search. We conducted a changepoint analysis (CPA) to detect significant statistical changes in the “naloxone” trend from 2004 to 2018. Cross-correlation function analyses were done to examine the correlation between 2 time series: year-wise relative search volume (RSV) for “naloxone” and “drug overdose” with the age-adjusted drug overdose mortality rate. Pearson’s correlation was performed for the state-wise age-adjusted mortality rate due to drug overdose and RSV for “naloxone” and “drug overdose.” <b><i>Results:</i></b> Smoothed and regressed GT of “naloxone” were similar to the “opioid overdose” trend published by the National Center for Health Statistics. The CPA showed 2 statistically significant points in 2011 and 2015. CPA of year-wise RSV for “naloxone” and “drug overdose” showed significantly positive correlation with the age-adjusted drug overdose mortality at lag zero. State-wise RSV for “naloxone” and “drug overdose” too showed a strong and significant positive correlation with the state-wise mortality data. <b><i>Discussion/Conclusion:</i></b> Inexpensive, publicly accessible, real-time GT data could supplement and strengthen the monitoring of opioid overdose epidemic if used in conjunction with the existing official data sources.