Influence of Mass Media on Italian Web Users During the COVID-19 Pandemic: Infodemiological Analysis

Abstract

Background: Concurrently with the COVID-19 pandemic, the world has been facing a growing infodemic, which has caused severe damage to economic and health systems and has often compromised the effectiveness of infection containment regulations. Although this infodemic has spread mainly through social media, there are numerous occasions on which mass media outlets have shared dangerous information, giving resonance to statements without a scientific basis. For these reasons, infoveillance and infodemiology methods are increasingly exploited to monitor information traffic on the web and make epidemiological predictions.

Objective: The purpose of this paper is to estimate the impact of Italian mass media on users' web searches to understand the role of press and television channels in both the infodemic and the interest of Italian netizens in COVID-19.

Methods: We collected the headlines published from January 2020 to March 2021 containing specific COVID-19-related keywords published on PubMed, Google, the Italian Ministry of Health website, and the most-read newspapers in Italy. We evaluated the percentages of infodemic terms on these platforms. Through Google Trends, we searched for cross-correlations between newspaper headlines and COVID-19-related web searches. Finally, we analyzed the web interest in infodemic content posted on YouTube.

Results: During the first wave of COVID-19, the Italian press preferred to draw on infodemic terms (rate of adoption: 1.6%-6.3%) and moderately infodemic terms (rate of adoption: 88%-94%), while scientific sources favored the correct names (rate of adoption: 65%-88%). The correlational analysis showed that the press heavily influenced users in adopting terms to identify the novel coronavirus (cross-correlations of ≥0.74 to ≤0.89, P value <.001; maximum lag=1 day). The use of scientific denominations by the press reached acceptable values only during the third wave (approximately 80%, except for the television services Rai and Mediaset). Web queries about COVID-19 symptoms also appeared to be influenced by the press (best average correlation=0.92, P<.007). Furthermore, web users showed pronounced interest in YouTube videos of an infodemic nature. Finally, the press gave resonance to serious "fake news" on COVID-19, which caused pronounced spikes of interest from web users.

Conclusions: Our results suggest that the Italian mass media have played a decisive role in spreading the COVID-19 infodemic and addressing netizens' web interest, thus favoring the adoption of terms that are unsuitable for identifying COVID-19. Therefore, the directors of news channels and newspapers should be more cautious, and government dissemination agencies should exert more control over such news stories.

Keywords: COVID-19; Google Trends; infodemic; infodemiology; infoveillance; mass media; mass media influence; media coverage; social media.

Figure 1

Comparison of the weekly RSVs of the keywords coronavirus (yellow) and covid (green) with the number of times the terms were adopted by the newspaper La Repubblica (blue and red, respectively) from January 1 to September 6, 2020. All values were normalized to 100. HDI: headlines daily increase; RSV: relative search volume.

Figure 2

Comparison between the daily RSVs (red) of the keyword coronavirus, new COVID-19 cases, and COVID-19 deaths, and the number of times they were adopted by the newspaper La Repubblica (blue) from January 11 to March 15, 2020. All values are normalized to 100. The black lines represent the beginning or the end of a trend. d.i.: daily increase; RSV: relative search volume.

Figure 3

Causal implications scheme and linear regression results. RSV: relative search volume; VE: variability explained.

Figure 4

Relative search volumes (RSVs) of infodemic queries since the start of the COVID-19 pandemic. The astrazeneca query is not shown to enable visualization of the other RSV trends.