Abstract
Early public awareness of multiple preventive measures was observed in regions with lower average number of daily cases. Our findings indicate optimal public health communication regarding masks, hand washing, hand sanitiser, and disinfectant in the general population during early stages of the COVID-19 pandemic. Early public awareness may facilitate future disease control efforts by public health authorities.
Introduction
In the early stages of the COVID-19 pandemic, before vaccines or proven pharmaceutical treatments were available, public behavior was the primary line of defense. However, measuring "public awareness" in real-time across the globe is notoriously difficult. This study addresses the gap in understanding how the timing of public interest in preventive measures (like masking and hand washing) relates to the actual progression of the disease. While previous studies looked at regional data, this paper takes a global approach to see if online search behavior can serve as a proxy for public health readiness and whether that readiness actually correlates with lower infection rates.
Methods
The researchers employed a retrospective "infodemiology" approach—the analysis of online search trends—using Google Trends (GT) data from January 1 to June 30, 2020. Instead of searching for specific strings, they used "search topics" for five categories: ‘mask’, ‘hand washing’, ‘social distancing’, ‘hand sanitizer’, and ‘disinfectant’. This allowed the study to capture searches across different languages and synonyms globally.
Relative Search Volume (RSV) data, normalized on a scale of 0–100, was analyzed alongside the average number of daily cases for 190 countries using Johns Hopkins University data. Pearson correlation coefficients were calculated to compare geographical RSV with case counts.
Global Search Trends
The temporal analysis of search volume revealed distinct peaks in public interest corresponding to major global events.
Most RSV peaks coincided with major WHO announcements, such as the declaration of a Public Health Emergency of International Concern (Jan 30) and the official pandemic declaration (March 11). Notably, 'mask' was the most dominant search query globally, exceeding interest in 'hand sanitizer' by more than threefold at its peak.
Key Insights and Results
The study identified a significant "Early Bird" effect: countries with lower average daily cases (e.g., Hong Kong, Taiwan, Vietnam) saw their search interest in preventive measures peak very early (late January to early February). In contrast, countries with high case counts (e.g., US, Brazil) saw peaks much later (late March to April). For instance, Hong Kong saw an early 'mask' peak on Jan 29 (Avg. daily cases: 7.75), while the United States saw a late 'mask' peak on April 4 (Avg. daily cases: 16,789.11).
Statistical analysis showed mild negative correlations between case counts and searches for ‘hand washing’ (r = -0.44) and ‘social distancing’ (r = -0.38). This implies that as search interest in these topics increased early on, average daily cases tended to be lower.
The study also highlighted the influence of public figures. A massive, 11-fold spike in "disinfectant" searches occurred on April 24, 2020, following a controversial speech by then-US President Donald Trump. This demonstrates how public figures can instantly shift global information-seeking behavior. Furthermore, interest in "staying home" (a colloquial term) peaked with a magnitude 50% higher than the technical term "social distancing," suggesting that public health messaging is more effective when using common language.
Limitations and Future Work
While search volume is a powerful proxy, it indicates interest rather than actual compliance or correct implementation of measures. Additionally, the reliance on Google Trends excludes regions like mainland China, where Baidu is the dominant search engine. Future research should integrate data from regional engines and social media platforms (e.g., Twitter, TikTok) to provide a more granular view. The study also did not account for confounding factors like government policy stringency, healthcare infrastructure, or population density, which also heavily influence case counts.
Summary updated