Here, I present an investigation into whether this change implemented by Google indeed had an impact on the GHT data retrieved and provide the first evidence that future research using the GHT platform should refrain from comparing data obtained from 1 January 2022 onwards to dates before 2022.Ģ.1. The GHT documentation has also not been updated yet to indicate that such a change was made, risking the potential that erroneous conclusions can be made in the future. Such changes impact ongoing research, especially when future research efforts seek to compare periods before and after the implementation of such a change. Google also indicated that any changes in search interest dating 1 January 2022 might be attributable to this change. The said changes were implemented on 18 July 2022, with all data from 1 January 2022 being altered to include this new comprehensive search request sample. Recently, Google announced via email that the GHT API “will be improved by providing higher precision responses by using a more comprehensive sample of search requests” ( Supplementary Email S1).

GHT was used to explore the search behavior of African Internet users related to the COVID-19 pandemic, as a prediction tool for dengue fever in Brazil, and to gauge interest in pre-exposure prophylaxis in the United States of America, with mixed reports on the effectiveness of this tool in infodemiology and epidemiology. The data retrieval period was set to the ‘past 90 days’ (blue line) and the ‘past 30 days’ (orange line).įor those interested in comparing search activity in different regions or time periods, Google offers limited access to the Google Health Trends (GHT) API. Google Trends results for ‘Coronavirus disease 2019′, retrieved on 12 September 2022 from the Google Trends platform ( ) based on Worldwide web searches. The extraction of Google Trends data can be automated, to some extent, using unofficial application programming interfaces (APIs), such as pytrends (v4.8.0) for Python. Since this is a scaled metric, based on the number of searches in the geographical limitation selected for the searched keyword, comparisons between regions are not possible with Google Trends data. For example, searching the same keyword for different time periods on the same geographical boundary yields different results ( Figure 1). Although this platform gives the user an indication of the dates or times that a specific phrase was searched the most frequently, it lacks the ability for users to compare results from different periods. Google Trends retrieves a relative search volume (RSV), a metric ranging from 0 to 100 and based on the proportional popularity of the keyword in a specific geographic region for the selected period. Specific to Google’s search engine, the unrestricted Google Trends platform (, accessed on 22 September 2022) is open for all to explore how specific demographics searched for certain keywords. Data extracted from (accessed on 4 June 2022) under Creative Commons Attribution-Share Alike 3.0 Unported License. Future research using the GHT API should not attribute increases in GHT data from 1 January 2022 onward as being reflective of increased search activity for a specific keyword, but rather attribute it to the implemented change to the GHT sampling strategy.ġ Market share as of June 2022. In general, the extracted values are higher after the implemented changes, compared to the values extracted before the change. Comparing 177 weekly probabilities for short search sessions of 421 Freebase IDs in thirty geographies extracted from GHT both before and after the implemented change, a low correlation (median of all Spearman ρ = 0.262 ) between these data was observed for the year 2022. Evidence is provided here for an initial exploration of how this change impacted the data obtained from GHT. Recently, Google implemented changes to the GHT platform. Tools such as Google Trends and Google Health Trends (GHT) can be used to gauge search activity in key geographical regions and for specific periods of time. Accessing such search activity allows infodemiologists a glimpse into the collective online mind. In an ever-increasingly online world, many Internet users seek information from online search engines such as Google.

Official google trends api