Online behavioural patterns for Coronavirus disease 2019 (COVID-19) in the United Kingdom

Google Trends as a predictive tool in the era of COVID-19: a scoping review

Google Trends has been extensively used in different sectors from finance to tourism, the economy, fashion, the fun industry, the oil trade, and healthcare. This scoping review aims to summarize the role of Google Trends as a monitoring and a predicting tool in the COVID-19 pandemic. Inclusion criteria for this scoping review were original English-language peer-reviewed research articles on the COVID-19 pandemic conducted in 2020 using Google Trends as a search tool. Articles that were in a language other than English, were only in abstract form, or did not discuss the role of Google Trends during the COVID-19 pandemic were excluded. According to these criteria, a total of 81 studies were included to cover the period of the first year after the emergence of the crisis. Google Trends can potentially help health authorities to plan and control pandemics earlier and to decrease the risk of infection among people.

Vector Autoregression for Forecasting the Number of COVID-19 Cases and Analyzing Behavioral Indicators in the Philippines: Ecologic Time-Trend Study

BACKGROUND

Traditional surveillance systems rely on routine collection of data. The inherent delay in retrieval and analysis of data leads to reactionary rather than preventive measures. Forecasting and analysis of behavior-related data can supplement the information from traditional surveillance systems.

OBJECTIVE

We assessed the use of behavioral indicators, such as the general public's interest in the risk of contracting SARS-CoV-2 and changes in their mobility, in building a vector autoregression model for forecasting and analysis of the relationships of these indicators with the number of COVID-19 cases in the National Capital Region.

METHODS

An etiologic, time-trend, ecologic study design was used to forecast the daily number of cases in 3 periods during the resurgence of COVID-19. We determined the lag length by combining knowledge on the epidemiology of SARS-CoV-2 and information criteria measures. We fitted 2 models to the training data set and computed their out-of-sample forecasts. Model 1 contains changes in mobility and number of cases with a dummy variable for the day of the week, while model 2 also includes the general public's interest. The forecast accuracy of the models was compared using mean absolute percentage error. Granger causality test was performed to determine whether changes in mobility and public's interest improved the prediction of cases. We tested the assumptions of the model through the Augmented Dickey-Fuller test, Lagrange multiplier test, and assessment of the moduli of eigenvalues.

RESULTS

A vector autoregression (8) model was fitted to the training data as the information criteria measures suggest the appropriateness of 8. Both models generated forecasts with similar trends to the actual number of cases during the forecast period of August 11-18 and September 15-22. However, the difference in the performance of the 2 models became substantial from January 28 to February 4, as the accuracy of model 2 remained within reasonable limits (mean absolute percentage error [MAPE]=21.4%) while model 1 became inaccurate (MAPE=74.2%). The results of the Granger causality test suggest that the relationship of public interest with number of cases changed over time. During the forecast period of August 11-18, only change in mobility (P=.002) improved the forecasting of cases, while public interest was also found to Granger-cause the number of cases during September 15-22 (P=.001) and January 28 to February 4 (P=.003).

CONCLUSIONS

To the best of our knowledge, this is the first study that forecasted the number of COVID-19 cases and explored the relationship of behavioral indicators with the number of COVID-19 cases in the Philippines. The resemblance of the forecasts from model 2 with the actual data suggests its potential in providing information about future contingencies. Granger causality also implies the importance of examining changes in mobility and public interest for surveillance purposes.

Spatiotemporal Variations of "Triple-demic" Outbreaks of Respiratory Infections in the United States in the Post-COVID-19 Era

Objectives

The United States confronted a "triple-demic" of influenza, respiratory syncytial virus, and COVID-19 in the winter of 2022, resulting in increased respiratory infections and a higher demand for medical supplies. It is urgent to analyze each epidemic and their co-occurrence in space and time to identify hotspots and provide insights for public health strategy.

Methods

We used retrospective space-time scan statistics to retrospect the situation of COVID-19, influenza, and RSV in 51 US states from October 2021 to February 2022, and then applied prospective space-time scan statistics to monitor spatiotemporal variations of each individual epidemic, respectively and collectively from October 2022 to February 2023.

Results

Our analysis indicated that compared to the winter of 2021, COVID-19 cases decreased while influenza and RSV infections increased significantly during the winter of 2022. We revealed that a twin-demic high-risk cluster of influenza and COVID-19 but no triple-demic clusters emerged during the winter of 2021. We further identified a large high-risk cluster of triple-demic in the central US from late November, with COVID-19, influenza, and RSV having relative risks of 1.14, 1.90, and 1.59, respectively. The number of states at high risk for multiple-demic increased from 15 in October 2022 to 21 in January 2023.

Conclusion

Our study provides a novel spatiotemporal perspective to explore and monitor the transmission patterns of the triple epidemic, which could inform public health authorities' resource allocation to mitigate future outbreaks.

Online information seeking during the COVID-19 pandemic: A cross-country analysis

The aim of this study was to investigate the coronavirus-related web-searching patterns of people from the 10 most affected nations in September 2020. The authors extracted all searches for the sample nations, consisting of the two words ‘COVID-19’ and ‘coronavirus’ and their variations, from Google Trends for the complete year of 2020. The results showed a discrepancy due to the priority of the language used during searches for coronavirus-related information. The time span of the attention level of citizens towards coronavirus-related information was relatively short (about one month). This supports the assumption of the activation model of information exposure that information which generates a negative affect is not welcomed by users. The findings have practical implications for governments and health authorities in, for example, launching information services for citizens in the early months of a pandemic and them remaining as the preferred source of information for citizens.

Forecasting and Surveillance of COVID-19 Spread Using Google Trends: Literature Review

The probability of future Coronavirus Disease (COVID)-19 waves remains high, thus COVID-19 surveillance and forecasting remains important. Online search engines harvest vast amounts of data from the general population in real time and make these data publicly accessible via such tools as Google Trends (GT). Therefore, the aim of this study was to review the literature about possible use of GT for COVID-19 surveillance and prediction of its outbreaks. We collected and reviewed articles about the possible use of GT for COVID-19 surveillance published in the first 2 years of the pandemic. We resulted in 54 publications that were used in this review. The majority of the studies (83.3%) included in this review showed positive results of the possible use of GT for forecasting COVID-19 outbreaks. Most of the studies were performed in English-speaking countries (61.1%). The most frequently used keyword was "coronavirus" (53.7%), followed by "COVID-19" (31.5%) and "COVID" (20.4%). Many authors have made analyses in multiple countries (46.3%) and obtained the same results for the majority of them, thus showing the robustness of the chosen methods. Various methods including long short-term memory (3.7%), random forest regression (3.7%), Adaboost algorithm (1.9%), autoregressive integrated moving average, neural network autoregression (1.9%), and vector error correction modeling (1.9%) were used for the analysis. It was seen that most of the publications with positive results (72.2%) were using data from the first wave of the COVID-19 pandemic. Later, the search volumes reduced even though the incidence peaked. In most countries, the use of GT data showed to be beneficial for forecasting and surveillance of COVID-19 spread.

eHealth Literacy of Chinese Residents During the Coronavirus Disease 2019 Pandemic: A Cross-sectional Survey

The Coronavirus Disease 2019 (COVID-19) pandemic has become a leading societal concern. eHealth literacy is important in the prevention and control of this pandemic. The purpose of this study is to identify eHealth literacy of Chinese residents about the COVID-19 pandemic and factors influencing eHealth literacy. A total of 15 694 individuals clicked on the link to the questionnaire, and 15 000 agreed to participate and completed the questionnaire for a response rate of 95.58%. Descriptive statistics, χ2 test, and logistic regression analysis were conducted to analyze participants' level of eHealth literacy about COVID-19 and its influencing factors. The results showed 52.2% of participants had relatively lower eHealth literacy regarding COVID-19 (eHealth literacy score ≤ 48). The scores of the information judgment dimension (3.09 ± 0.71) and information utilization dimension (3.18 ± 0.67) of the eHealth literacy scale were relatively lower. The logistics regression showed that sex, age, education level, level of uncertainty, having people around the respondent diagnosed with COVID-19, relationship with family, and relationship with others were associated to eHealth literacy (χ2 = 969.135, P < .001). The public's eHealth literacy about COVID-19 needs to be improved, especially the ability to judge and utilize online information. Close collaboration among global health agencies, governments, healthcare institutions, and media is needed to provide reliable online information to the public. Interventions to improve eHealth literacy should take into account and accentuate the importance of sex, age, educational background, level of uncertainty, exposure to disease, and social support.

Cohort profile: The UK COVID-19 Public Experiences (COPE) prospective longitudinal mixed-methods study of health and well-being during the SARSCoV2 coronavirus pandemic

Public perceptions of pandemic viral threats and government policies can influence adherence to containment, delay, and mitigation policies such as physical distancing, hygienic practices, use of physical barriers, uptake of testing, contact tracing, and vaccination programs. The UK COVID-19 Public Experiences (COPE) study aims to identify determinants of health behaviour using the Capability, Opportunity, Motivation (COM-B) model using a longitudinal mixed-methods approach. Here, we provide a detailed description of the demographic and self-reported health characteristics of the COPE cohort at baseline assessment, an overview of data collected, and plans for follow-up of the cohort. The COPE baseline survey was completed by 11,113 UK adult residents (18+ years of age). Baseline data collection started on the 13th of March 2020 (10-days before the introduction of the first national COVID-19 lockdown in the UK) and finished on the 13th of April 2020. Participants were recruited via the HealthWise Wales (HWW) research registry and through social media snowballing and advertising (Facebook®, Twitter®, Instagram®). Participants were predominantly female (69%), over 50 years of age (68%), identified as white (98%), and were living with their partner (68%). A large proportion (67%) had a college/university level education, and half reported a pre-existing health condition (50%). Initial follow-up plans for the cohort included in-depth surveys at 3-months and 12-months after the first UK national lockdown to assess short and medium-term effects of the pandemic on health behaviour and subjective health and well-being. Additional consent will be sought from participants at follow-up for data linkage and surveys at 18 and 24-months after the initial UK national lockdown. A large non-random sample was recruited to the COPE cohort during the early stages of the COVID-19 pandemic, which will enable longitudinal analysis of the determinants of health behaviour and changes in subjective health and well-being over the course of the pandemic.

COVID-19 pandemic in the Middle East countries: coronavirus-seeking behavior versus coronavirus-related publications

The spread of COVID-19 has created a fundamental need for coordinated mechanisms responding to outbreaks in different sectors. One of the main sectors relates to information supply and demand in the middle of this pandemic in the digital environment. It could be called an infodemiology. It is known as a promising approach to solving the challenge in the present age. At this level, the purpose of this article is to investigate the COVID-19 related search process by field research. Data were retrieved from Google Trends in Middle Eastern countries alongside scientific research output of Middle Eastern scientists towards COVID-19 in Web of Science, Scopus, and PubMed. Daily COVID-19 cases and deaths were retrieved from the World Health Organization. We searched for descriptive statistical analyses to detect coronavirus-seeking behavior versus coronavirus releases in the Middle East in 2020. Findings show that people in the Middle East use various keyword solutions to search for COVID-19 in Google. There is a significant correlation between coronavirus confirmed cases and scientific productivity (January 2020-December 2020). Also, there is a positive association between the number of deaths and the number of scientific publications (except Jordan). It was a positive and significant association between online coronavirus-seeking behavior on Google (RSVs) and the confirmed cases (except Syria and Yemen). Furthermore, it was a positive relationship between RSVs and scientific productivity in the Middle East (except Bahrain and Qatar). From an infodemiological viewpoint, there is a significant correlation between coronavirus information demand and its information provision.

The Influence of Media Coverage and Governmental Policies on Google Queries Related to COVID-19 Cutaneous Symptoms: Infodemiology Study

Background

During COVID-19, studies have reported the appearance of internet searches for disease symptoms before their validation by the World Health Organization. This suggested that monitoring of these searches with tools including Google Trends may help monitor the pandemic itself. In Europe and North America, dermatologists reported an unexpected outbreak of cutaneous acral lesions (eg, chilblain-like lesions) in April 2020. However, external factors such as public communications may also hinder the use of Google Trends as an infodemiology tool.

Objective

The study aimed to assess the impact of media announcements and lockdown enforcement on internet searches related to cutaneous acral lesions during the COVID-19 outbreak in 2020.

Methods

Two searches on Google Trends, including daily relative search volumes for (1) “toe” or “chilblains” and (2) “coronavirus,” were performed from January 1 to May 16, 2020, with the United States, the United Kingdom, France, Italy, Spain, and Germany as the countries of choice. The ratio of interest over time in “chilblains” and “coronavirus” was plotted. To assess the impact of lockdown enforcement and media coverage on these internet searches, we performed an interrupted time-series analysis for each country.

Results

The ratio of interest over time in “chilblains” to “coronavirus” showed a constant upward trend. In France, Italy, and the United Kingdom, lockdown enforcement was associated with a significant slope change for “chilblain” searches with a variation coefficient of 1.06 (SE 0.42) (P=0.01), 1.04 (SE 0.28) (P<.01), and 1.21 (SE 0.44) (P=0.01), respectively. After media announcements, these ratios significantly increased in France, Spain, Italy, and the United States with variation coefficients of 18.95 (SE 5.77) (P=.001), 31.31 (SE 6.31) (P<.001), 14.57 (SE 6.33) (P=.02), and 11.24 (SE 4.93) (P=.02), respectively, followed by a significant downward trend in France (–1.82 [SE 0.45]), Spain (–1.10 [SE 0.38]), and Italy (–0.93 [SE 0.33]) (P<.001, P=0.004, and P<.001, respectively). The adjusted R² values were 0.311, 0.351, 0.325, and 0.305 for France, Spain, Italy, and the United States, respectively, suggesting an average correlation between time and the search volume; however, this correlation was weak for Germany and the United Kingdom.

Conclusions

To date, the association between chilblain-like lesions and COVID-19 remains controversial; however, our results indicate that Google queries of “chilblain” were highly influenced by media coverage and government policies, indicating that caution should be exercised when using Google Trends as a monitoring tool for emerging diseases.

Google Trends Data and COVID-19 in Europe: Correlations and model enhancement are European wide

The current COVID-19 pandemic offers a unique opportunity to examine the utility of Internet search data in disease modelling across multiple countries. Most such studies typically examine trends within only a single country, with few going beyond describing the relationship between search data patterns and disease occurrence. Google Trends data (GTD) indicating the volume of Internet searching on 'coronavirus' were obtained for a range of European countries along with corresponding incident case numbers. Significant positive correlations between GTD with incident case numbers occurred across European countries, with the strongest correlations being obtained using contemporaneous data for most countries. GTD was then integrated into a distributed lag model; this improved model quality for both the increasing and decreasing epidemic phases. These results show the utility of Internet search data in disease modelling, with possible implications for cross country analysis.