Background: Codeine, the most widely used opioid in the world, is often prescribed to suppress mild to moderate pain and cough but carries risks of abuse. Growing concerns about the potential for abuse led the French Ministry of Health to change codeine to a prescription-only drug on July 12, 2017.
Objectives: We sought to assess the temporal and spatial trends of searches for "codeine" in the general population before and after implementing the risk minimization measure (RMM) in France. We also assessed parallel trends for other opioids without intervention in France and other countries without intervention.
Methods: We reanalyzed publicly available online data from Google Trends (GT), which was previously used in health informatics to study human behaviour. The primary outcome measure, the rate of public engagement, was defined as the number of searches for codeine relative to the total number of searches at multiple equidistant monthly periods in the selected geographies. The source population was France, but a parallel assessment was made in Germany, the UK and worldwide. In France, in addition to "codeine," we used other search terms for popular purple drank (street mixture of codeine and alcohol), Izalgi (over the contour combination of paracetamol and opium), and tramadol (synthetic opioids). We reported the results descriptively and visually using correlation values in an interrupted time series approach.
Results: Public interest in codeine in France from July 2015 to July 2017, i.e., two years before the RMM, was on the rise (r=.442). The trend, however, changed direction after the RMM and gradually decreased in the following two years (r=-.271). We also found a similar downward trend occurred for the popular street mixed of codeine and alcohol in France two years after the intervention (r= -.479). However, we found no pattern were found for izalgi (over the contour combination of paracetamol and opium) and tramadol in France or Germany products and tramadol were unaffected in France following the intervention. Repeating the methodology for izalgi and tramadol in France and in Germany and the UK without the RMM revealed no change after the French RMM.
Conclusions: This study is subject to limitations, such as its retrospective nature based on online searches of selected countries, and its inherent margin of error could bias the correlation results., It showed that French safety RMM successfully reduced public engagement rates with abuse-prone codeine and the high-profile illicit mixed codeine and alcohol, purple drank. Alongside our previous work on zolpidem, Google Trends could help assess safety measures at the population level; however, further investigation is needed to better understand advantages and limitations of this source of information.
