Expanded Limitations and Recommendations for Using Google Trends for Public Interest and Psychiatric Research

Article information

Psychiatry Investig. 2023;20(8):786-787

Publication date (electronic) : 2023 August 11

doi : https://doi.org/10.30773/pi.2023.0078

Rowalt Alibudbud

Department of Sociology and Behavioral Sciences, College of Liberal Arts, De La Salle University, Manila, Philippines

Correspondence: Rowalt Alibudbud, MD Department of Sociology and Behavioral Sciences, College of Liberal Arts, De La Salle University, 2401 Taft Avenue, Manila 1004, Philippines Tel: +63-925-576-9258, Fax: +63-302-4234 E-mail: rowalt.alibudbud@dlsu.edu.ph

Received 2023 March 7; Revised 2023 April 10; Accepted 2023 June 5.

Dear Editor,

In a recent paper published in Psychiatry Investigation, Becerra-García et al. [1] excellently explored the worldwide public interest regarding psychiatric treatments, including psychotherapy, antidepressant, antipsychotic, and anxiolytic, from March 2020 to December 2021 using Google Trends. The authors found that interest in antidepressants and antipsychotics increased between 2020 and 2021 [1]. Although their study judiciously expounded the limits of their research [1], they may be constrained by the length limitations of their article type to systematically delve into the different limitations and recommendations when analyzing data from Google Trends. Thus, this letter expands on several limitations and recommendations that may also deserve mention in studies utilizing Google Trends, such as the study of Becerra-García et al. [1], as well as other infodemiological studies about psychiatric and mental health phenomena [1-4].

The use of Google Trends for health research and public health is subsumed into the field of infodemiology, a portmanteau of information and epidemiology [2]. Infodemiology is defined as the science and study of information in an online medium to inform public health and health policy [2]. Aside from psychiatry, Google Trends and, in general, infodemiology, have been utilized to explore and understand public interests and prevalence rates of various diseases and disorders, such as flu, HPV, cancer, and smoking [2-5].

Google Trends may only reflect the online interest of those who utilize the Google search engine [1-5]. Therefore, further exploration of public interest can be done by exploring online interest in areas with low Google market share using data from other search engines, such as China using the Baidu index (https://www.baidu.com) [2,4,5]. Second, Google Trends may have limited reliability in areas with low freedom of speech [2,4,5]. Thus, utilizing traditional methods, such as surveys, may be more appropriate in regions with restricted freedom of speech [2]. Third, it was also highlighted that spelling choices could affect the results [1]. Therefore, future research can utilize “topic” keywords, which are groups of terms in Google Trends that share the same concept in any language [2,3]. Fourth, while it was emphasized that there is no standard procedure for data collection,1 previous reviews highlighted that reporting standards could be used to promote the replicability of studies done in Google Trends [2,4,5]. For instance, it was suggested that the selected keywords, period, region, query category, and date of retrieval be documented in the methodology of research utilizing Google Trends [2,4,5].

Fifth, reviews reveal that the correlation between Google Trends and epidemiologic data can vary [2,4,5]. For example, there can be a low correlation between Google Trends and epidemiological data [4]. Therefore, while it was alluded that the increased public interest in antidepressants and antipsychotics may reflect a potential increase in global mental health problems [1], it may be necessary to conduct further analysis and explore additional factors to ascertain the correlation between Google Trends and epidemiologic data. Moreover, other sources of data from the internet can be used to supplement Google Trends, such as Facebook, Wikipedia, and Twitter data [2], in exploring and understanding public interests and prevalence rates of various diseases and disorders. Based on the above-mentioned, evidence suggests that Google Trends remains a surrogate measure for public interest [2-5]. Therefore, traditional measures, such as surveys, can be conducted to comprehensively understand public interest in psychiatric treatments.