Web-Based Content on Diet and Nutrition Written in Japanese: Infodemiology Study Based on Google Trends and Google Search

Abstract

Background: The increased availability of content of uncertain integrity obtained through the internet is a major concern. To date, however, there has been no comprehensive scrutiny of the fitness-for-purpose of web-based content on diet and nutrition.

Objective: This cross-sectional study aims to describe diet- and nutrition-related web-based content written in Japanese, identified via a systematic extraction strategy using Google Trends and Google Search.

Methods: We first identified keywords relevant for extracting web-based content (eg, blogs) on diet and nutrition written in Japanese using Google Trends. This process included identification of 638 seed terms, identification of approximately 1500 pairs of related queries (top) and search terms, the top 10% of which were extracted to identify 160 relevant pairs of related queries (top) and search terms, and identification of 107 keywords for search. We then extracted relevant web-based content using Google Search.

Results: The content (N=1703) examined here was extracted following a search based on 107 keywords. The most common themes included food and beverages (390/1703, 22.9%), weight management (366/1703, 21.49%), health benefits (261/1703, 15.33%), and healthy eating (235/1703, 13.8%). The main disseminators were information technology companies and mass media (474/1703, 27.83%), food manufacturers (246/1703, 14.45%), other (236/1703, 13.86%), and medical institutions (214/1703, 12.57%). Less than half of the content (790/1703, 46.39%) clearly indicated the involvement of editors or writers. More than half of the content (983/1703, 57.72%) was accompanied by one or more types of advertisement. The proportion of content with any type of citation reference was 40.05% (682/1703). The themes and disseminators of content were significantly associated with the involvement of editors or writers, accompaniment with advertisement, and citation of reference. In particular, content focusing on weight management was more likely to clearly indicate the involvement of editors or writers (212/366, 57.9%) and to be accompanied by advertisement (273/366, 74.6%), but less likely to have references cited (128/366, 35%). Content from medical institutions was less likely to have citation references (62/214, 29%).

Conclusions: This study highlights concerns regarding the authorship, conflicts of interest (advertising), and the scientific credibility of web-based diet- and nutrition-related information written in Japanese. Nutrition professionals and experts should take these findings seriously because exposure to nutritional information that lacks context or seems contradictory can lead to confusion and backlash among consumers. However, more research is needed to draw firm conclusions about the accuracy and quality of web-based diet- and nutrition-related content and whether similar results can be obtained in other major mass media or social media outlets and even other languages.

Keywords: Japanese language; diet; information; internet; nutrition; web.

Figure 1

Flowchart of a systematic process for identifying keywords relevant for extracting web-based content on diet and nutrition written in Japanese. (A) identification of 638 seed terms; (B) identification of about 15,000 pairs of search terms and related queries (top); (C) identification of 160 relevant pairs of related queries (top) and search terms; (D) identification of 107 keywords for search.

Figure 2

Screenshot of Google Trends as an example. In this study, all 638 seed terms were used as search terms (A and B). We conducted a search for each search term (A: dieting in this example) in comparison with the criterion search term meaning meals (B); in other words, only 2 search terms were compared at once. All searches were carried out under uniform search conditions with regard to region, time period, category, and field (C). By doing this, we obtained the relative search volume (RSV; an index of search volume adjusted to the number of Google users in a given geographic area and period, ranging from 0 to 100; (D) relative to the one criterion search term (B: meals) for all search terms (A: dieting in this example).

Figure 3

Screenshot of Google Trends as an example. The search conducted using Google Trends (as shown in Figure 2) simultaneously provides data on related queries (top), namely the terms that are most frequently searched with the search term entered in the same search session, within the chosen category, country, or region. As each related query (top) has a value of relative search volume (RSV), this value was multiplied by the RSV value of the originating search term (adjusted for the RSV value of the criterion search term), creating a list of pairs of related queries (top) and search terms according to their popularity.

Figure 4

Search terms (outer layer) used to identify web-based content related to diet and nutrition in Japanese and the topic themes (inner layer). The number of web-based content is shown (N=1703).

Figure 5

Disseminators (outer layer) and their categories (inner layer) of web-based content related to diet and nutrition written in Japanese. The number of web-based content is shown (N=1703).