Infodemiology (ie, information epidemiology) is a field in health informatics defined as “the science of distribution and determinants of information in an electronic medium, specifically the Internet, or in a population, with the ultimate aim to inform public health and public policy” [1]. The first official mention of infodemiology according to a search in PubMed (ie, baring the term on the title) was by Gunther Eysenbach in 2002 [2]. However, infodemiology studies (ie, assessment of health-related topics using web-based data [3]) can be traced all the way back to 1996. Two more studies in the 2000s use the term (PubMed): one in 2004, where the quality of hospitals’ websites was assessed [4], and one in 2006, showing that flu data from Google correlated with influenza cases [5].

The large corpus of publications in infodemiology were present after 2009, with the first complete presentation and assessment of the subject being found in the scoping review of Bernardo et al [6] published with JMIR Publications—the main publisher of infodemiology and infoveillance studies.

Social media and search queries are the most popular sources for retrieving information from web-based sources. The use of social media is constantly expanding [7], with more users and including more features. Search query data is also of significant value, as they take into account the revealed and not the stated users’ preferences [8,9], but methodology should be designed with caution to ensure the validity of the results [10].

Popular social media data sources in infodemiology include Twitter [11-17], Facebook [18-22], and Instagram [23,24]. Queries from search engines are mostly retrieved by Google Trends [25-32], as well as Yandex [33-35], Baidu [36,37], Bing [38], Yahoo [39], and Daum [40,41]. Other popular sources include websites and platforms [42-45]; blogs, forums, and online communities [46-52]; and, what has received attention lately, mobile apps of certain health categories (eg, asthma [53] and heart failure self-care management [54]). Significant focus has been shown in combining two or more sources such as Facebook and Instagram [55], Facebook and Twitter Posts [56], US newspaper media and Facebook [57], and Google and Wikipedia [58].

The use of web-based sources offers an assessment of real-time information, whether it is from Twitter, Google, or other social media and search queries. For health data retrieved through traditional methods such as registries, questionnaires, or surveys, analysis and assessment can take time to perform. Thus, nowcasting using said methods is not trivial. However, web-based (real time) data has been shown to significantly contribute to the analysis and forecasting of certain diseases, outbreaks, and epidemics.

Therefore, this specific part of health informatics has benefitted from infodemiology. In particular, one of the most studied diseases is influenza, and several data sources have been employed to predict and assess flu-related topics [39,40,59-76]. Epidemics and infectious diseases that have been analyzed and assessed using infodemiology and infoveillance approaches include HIV/AIDS [77-79], measles [80-83], and the Zika virus [84-87].

Infodemiology topics have also been the subject of research for several reviews on various topics like curable sexually transmitted diseases (STDs) [88] and mental health disorders [89], and for data sources like search queries, social media [6], mobile phone apps [90], Twitter [91], and Google Trends [92].

Infodemiology has contributed to health assessment with the analysis of a range of topics. In specific, popular categories in the field of infodemiology and infoveillance include drugs [39,93,94], marijuana [95-97], depression and suicide [98-108], smoking and tobacco [109-116], electronic cigarettes (e-cigarettes) [117-126], and hookahs [127-130]. As far as chronic diseases are concerned, the assessment of diabetes [131-136] and multiple sclerosis [137,138] has benefitted from infodemiology and infoveillance studies. Other topics include breast cancer [139-142]; fitness and diet [143-146]; health care performance, evaluation, and dissemination [147,148]; and human papillomavirus (HPV) [149-154].

This review aims to update and expand the 2013 scoping review of Bernardo et al [6]. The authors of said review provided a well-structured outline of how infodemiology was employed in health informatics research up to that point, but as is evident, the large corpus of studies in the field have been published within the last couple of years. This update on the subject is important in identifying how infodemiology has contributed to health informatics over the past decade compared with traditional surveillance methods, the main web sources used, and the individual health categories and topics that have been explored. Apart from identifying the “metrics” of infodemiology studies (ie, numer of publications, thematic topics, journals and publishers, and data sources), this review aims to identify the advantages, disadvantages, and value and validity of infodemiology and infoveillance.

To select the publications in the fields of infodemiology and infoveillance, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Extension for Scoping Reviews [155,156] were followed. The procedure is depicted in the PRISMA flow diagram in Figure 1.

In the JMIR Publications database, all papers from the two relevant electronic collections (e-collections) were retrieved: 225 documents from the “Infodemiology and Infoveillance” [157] e-collection in the Journal of Medical Internet Research and 185 from the “Infodemiology, Infoveillance, and Digital Disease Surveillance” [158] e-collection in JMIR Public Health and Surveillance. After removing the duplicates and 2 documents for article type eligibility, 227 documents were extracted in total. Next, 66 documents were handpicked from the JMIR publications database based on searches of data sources (ie, “Twitter,” “Google Trends,” and “Google Flu Trends”). After 10 documents were excluded based on article type, a total of 56 documents were handpicked from the JMIR database.

Next, the term “infodemiology” was searched for in the PubMed database from January 1, 2009, to December 31, 2018, in the search field “Title-Abstract.” The search returned 94 documents. The term “infoveillance” was then entered for the same period and in the same field, and the search returned 51 documents. The duplicates were 39 in total, and 2 documents were excluded based on article type; a total of 104 documents were extracted from the PubMed database. The terms “infodemiology” and “infoveillance” were then independently searched for in the Scopus database in the “Article title, Abstract, Keywords” field for the full years 2009-2018. The search returned 88 and 43 documents, respectively (ie, a total of 131). After removing 33 duplicates and 4 documents for article type criteria, a total of 94 documents were extracted from the Scopus database.

Overall, all studies that included the terms “infodemiology” or “infoveillance” in the “Title/Abstract” field in PubMed up to December 2018 and all studies including the terms “infodemiology” or “infoveillance” in the “Article title, Abstract, Keywords” field up to 2018 in Scopus were selected. For JMIR, all articles in the two relevant e-collections, as well as the articles derived by the individual data source searches, were included in this review. Articles were only excluded based on article type eligibility (eg, correction, erratum). After removing the duplicates from the JMIR, PubMed, and Scopus databases, the total extracted documents from all databases were 338.

Table A1 in Multimedia Appendix 1 consists of the 338 selected publications included in this review, and shows the authors’ names, publication year, the title, and the journal used for the analysis. Figure 2 depicts the number of publications by year from 2009 to 2018.

The number of publications in the subject is increasing every year, with 2018 featuring more than one-fourth of the 338 total publications (n=89, 26.3%), and the publications from 2017 and 2018 combined accounted for more than half (n=171, 50.6%) of the total number of publications in the past decade.

The 338 extracted studies were published in 57 journals in total. The vast majority of the studies in infodemiology and infoveillance were published with JMIR Publications (n=282, 83.4%). Specifically, the Journal of Medical Internet Research features almost half of the publications (n=168, 49.7%), and JMIR Public Health and Surveillance features almost one-fourth of the examined studies (n=74, 21.9%).

Figure 3 consists of the numbers of publications per journal with >2 publications on the subject.

Journals that have published more than 1 paper in the subject included JMIR Mental Health (n=9), JMIR Research Protocols (n=7), JMIR Diabetes (n=4), JMIR Cancer (n=3), JMIR Medical Informatics (n=3), JMIR Pediatrics and Parenting (n=3), the Journal of Big Data (n=2), and the Interactive Journal of Medical Research (n=2).

Table A2 of Multimedia Appendix 1 consists of the 338 publications categorized by data source employed. Figure 4 depicts the popularity of the examined data sources in terms of number of publications (some publications employed more than one data source).

Of the 338 publications, the most popular source was Twitter with 45.0% (n=152) and is continuously gaining popularity. Google sources were in second place with 24.6% (n=83), followed by websites and platforms with 13.9% (n=47), blogs and forums with 10.1% (n=34), Facebook with 8.9% (n=30), and other search engines with 5.6% (n=19). The Google category consisted mainly of publications using Google Trends; although, the following Google tools have also been identified as main data sources in several publications: Google Flu Trends (n=6), Google Analytics (n=2), Google Insights (n=2), Google Correlate (n=1), Google Health (n=1), Google News (n=1), Google AdWords (n=1), Google Video (n=1), and Google Blog Search (n=1).

The “other search engines” category consists of Bing (n=7), Baidu (n=4), Yandex (n=4), Daum (n=2), and Yahoo (n=3), and the “other social media” category consists of YouTube (n=5), Yelp (n=5), Google+ (n=4), Foursquare (n=1), SoundCloud (n=1), Tumblr (n=1), Pinterest (n=1), and MySpace (n=1). Yahoo answers (n=2) was included in the blogs, forums, and communities category.

Although many health topics have been examined in infodemiology and infoveillance, some are significantly more popular. Figure 5 depicts the general categories in terms of number of publications, while Figure A1 of Multimedia Appendix 2 consists of the pie charts of their subcategories. All individual topics and subtopics identified in this review by number of publications can be found in Table A1 of Multimedia Appendix 2.

In the 338 publications examined in this review, the most popular subjects were conditions and diseases with 17.2% (n=58) and epidemics and outbreaks with 15.7% (n=53), followed by health care with 11.5% (n=39), drugs with 10.4% (n=35), smoking and alcohol with 8.6% (n=29), and mental health with 8.3% (n=28). Infectious diseases with 8.0% (n=27) and cancer with 6.8% (n=23) were also featured in several publications. The categories of diet and fitness with 4.1% (n=14) and mother and child with 2.7% (9) were less popular.