Content area
Objectives
Dengue fever is a major public health problem in countries like India, where traditional surveillance systems often suffer from delays. The study aims to examine the relationship between Google Trends data and the official record of dengue outbreaks in India as a supplementary tool to regular surveillance methods.
Methods
We used the Google Trends website to obtain the Google Trends data for the search terms “dengue fever,” “dengue symptoms,” and “dengue treatment” for the year 2023, along with the official record of the number of dengue outbreaks in the year 2023 from the Integrated Disease Surveillance Program (IDSP) website. Pearson’s correlation analysis, smoothed moving average, and the Toda-Yamamoto causality test were used to explore the strength, direction, and causality between the Google Trends data and official reports of the number of dengue outbreaks in India.
Results
The Toda-Yamamoto causality test revealed significant Granger causality between the search terms “dengue fever (p < 0.001),” “dengue symptoms (p < 0.001),” and “dengue treatment (p < 0.001)” with official records of the number of dengue outbreaks in India.
Conclusion
Google Trends data for the searched terms can supplement traditional surveillance methods for dengue outbreaks in India. Strong correlation coupled with significant Granger causality indicates its potential use as an early warning signal for dengue outbreaks in the country.
Introduction
Dengue fever is an endemic disease in tropical and subtropical countries such as India, presenting a considerable challenge to the public health system. Estimates indicate that the annual economic cost of dengue fever in India is approximately 1.1 billion US dollars. The disease results from an arbovirus that is transmitted by Aedes mosquitoes, which reproduce in artificial water collections such as coolers and plant pots. According to the World Health Organization (WHO), an estimated 390 million dengue infections occur every year, with severe outbreaks common in densely populated urban areas [1, 2].
In India in the year 2023, a total of 2.89 lakh cases and 485 deaths due to dengue were reported. There are recurrent dengue outbreaks during the monsoon season in India due to water collection, a breeding ground for Aedes mosquitoes, resulting in an increased incidence of dengue fever. Traditional surveillance methods for tracking dengue cases are used in many countries, such as India [3]. The dengue fever surveillance in India mainly relies on passive surveillance, where cases are reported from healthcare facilities to higher authorities. The Integrated Disease Surveillance Programme (IDSP) collects and provides weekly outbreak reports from both private and public healthcare facilities. It follows a decentralized system where the data from local health sites is first compiled at the district surveillance unit (DSU); from each district’s reports, it goes to the state surveillance unit (SSU), and finally, the report is consolidated at the national surveillance unit. But the reporting often suffers from delays [4]. In recent years, many advancements in digital surveillance tools have taken place, especially those based on big data and internet platforms [5]. Google Trends offers one such platform. It provides real-time insight into public interest in various health issues, like influenza, measles, and dengue [3, 6, 7, 8–9]. This data can supplement the traditional disease surveillance system. Studies done in the past have demonstrated the potential of Google Trends data in predicting influenza outbreaks, raising the possibility that similar methods might apply to dengue in India [10].
This study aims to analyze the relationship between Google Trends data for the selected search terms and the number of dengue outbreaks in India. It will explore the temporal correlation between Google Trends data and the official data of dengue outbreaks in India and will investigate the potential causality. By better understanding these dynamics, we hope to explore the potential of Google Trends data in dengue surveillance.
Materials and methods
Source of data
The Google Trends data for the search terms and the number of dengue outbreaks in India for the year 2023 was obtained. Three search terms were selected: “dengue fever,” “dengue symptoms,” and “dengue treatment,” respectively. Google Trends data for the selected search terms with weekly granularity was obtained from the Google Trends website [11]. The Google Trends data indicates relative interest among the public for the search terms, the value of which ranges from 0 to 100. The number of outbreaks of dengue in India for the year 2023 was obtained from the integrated disease surveillance platform (IDSP). The IDSP provides a weekly outbreak report [12]. Both the data are available freely in the public domain.
Statistical analysis
Both datasets were entered in MS Excel in a time series format with weekly granularity. The number of dengue outbreaks was normalized to make it comparable with the Google Trends data. For this, each data point was divided by the maximum value in the dataset, after which each obtained value was multiplied by 100. This scaled the value of dengue outbreak data, with a range from zero to 100, making both datasets on the same scale. The data was then uploaded in R software version 4.3.3 [13]. Pearson’s correlation coefficient was calculated as part of an exploratory analysis to see the direction and strength of the correlation between the search terms “dengue fever,” “dengue symptoms,” and “dengue treatment” and the normalized values of the number of dengue outbreaks in India. To visually interpret the relation, a line plot was also made between the search terms Google Trends data and the normalized value of the number of dengue outbreaks in India. A 3-week smoothed moving average curve was also plotted to remove minor fluctuations.
The Toda-Yamamoto causality test was then used to understand whether Google Trends search term data Granger causes the number of dengue outbreaks in India. Before applying the test, both the time series data for the search terms and normalized values of dengue outbreaks in India were assessed for stationarity. Datasets related to the search term “dengue treatment” and the normalized value of the number of dengue outbreaks in India were found to be non-stationary. Three different combinations of the Google Trends search terms and normalized values of the number of dengue outbreaks in India were made. For each combination, the maximum differencing required (dmax) was calculated and used for the Toda-Yamamoto causality test. Toda-Yamamoto causality tests do not require assumptions of stationarity and co-integration to be fulfilled for applying the test. Figure 1 depicts the entire process in a flow diagram.
[See PDF for image]
Fig. 1
Flow diagram depicting the methodological steps for analyzing the relationship between Google Trends data for the search term and normalized value of dengue outbreaks in India
Results
A positive and strong correlation was found between the Google Trends data for the search terms “dengue fever,” “dengue symptoms,” and “dengue treatment” and the normalized value of the number of dengue outbreaks in India.
Table 1. Correlation between Google trends of searched terms and normalized values of number of dengue outbreaks in India
Searched term | Pearson correlation coefficient | 95% confidence interval | p value |
|---|---|---|---|
Dengue fever | 0.82 | 0.70–0.89 | < 0.001 |
Dengue symptoms | 0.80 | 0.67–0.88 | < 0.001 |
Dengue treatment | 0.83 | 0.73–0.90 | < 0.001 |
To visually inspect the correlation, a line plot was made between the Google Trends data of the search terms and the normalized values of the number of dengue outbreaks in India obtained through the integrated disease surveillance platform. To remove minor fluctuations, a 3-week smoothed moving average curve is also plotted in the same graph.
[See PDF for image]
Fig. 2
Temporal comparison of dengue outbreaks in India (2023) using surveillance data and Google Trends data for the searched term “dengue fever”. * Surveillance data: from IDSP; Google trends data: Searched term Google trends data; MA: Moving average
[See PDF for image]
Fig. 3
Temporal comparison of dengue outbreaks in India (2023) using surveillance data and Google Trends data for the searched term “dengue symptoms”. * Surveillance data: from IDSP; Google trends data: Searched term Google trends data; MA: Moving average
[See PDF for image]
Fig. 4
Temporal comparison of dengue outbreaks in India (2023) using surveillance data and Google Trends data for the searched term “dengue treatment”. * Surveillance data: from IDSP; Google trends data: Searched term Google trends data; MA: Moving average
Causality test using the Toda-Yamamoto test
A Toda-Yomamato causality test was used to investigate whether Google Trends data for the search terms “dengue fever,” “dengue symptoms,” and “dengue treatment” Granger-causes the normalized values of the official number of dengue outbreaks in India. This test does not require the assumption of stationarity and co-integration. It is based on a vector autoregression (VAR) model that uses a modified Wald test.
Before applying the Toda-Yamamoto causality test, the stationarity of the time series data was checked using the augmented Dickey-Fuller test. The time series data of Google Trends for the searched terms “dengue fever” (Dickey-Fuller: -3.83; lag order 3; p 0.02) and “dengue symptoms” (Dickey-Fuller: -3.55; lag order 3; p 0.04) were found to be stationary. The time series data of the searched terms “dengue treatment” and the normalized value of the “official number of dengue outbreaks” were found to be non-stationary. Given this non-stationarity, the time series data of the Google Trends data of the search term “dengue treatment” and the normalized value of the official number of dengue outbreaks in India in the year 2023 were differenced. The third-order differenced Google Trends data of the search term “dengue treatment became stationary with a lag of 3 (Dickey-Fuller: -6.61; lag order 3; p 0.01). The first difference time series data of the normalized value of the official number of dengue outbreaks in India in the year 2023 became stationary with a lag of 3 (Dickey-Fuller: -4.87; lag order 3; p 0.01). No cointegration was found between the Google Trends data of the search terms “dengue fever,” “dengue symptoms,” and “dengue treatment” and normalized values of the “official number of dengue outbreaks in India in the year 2023,” which confirmed the appropriateness of the Toda-Yomamato test without requiring stationarity. The level of differencing required to make the series stationary provided the value of dmax.
Optimal lag length selection for the vector autoregressive (VAR) model is a critical requirement for the Toda-Yamamoto test. The lag length was selected based on the Akaike information criterion (AIC), which indicated three different lag lengths for the three different combinations of search terms in Google Trends data and the normalized value of the official number of dengue outbreaks in India: 9 lags for “dengue fever,” 5 lags for “dengue symptoms,” and 10 lags for “dengue treatment” search terms and the normalized value of the official number of dengue outbreaks in India, respectively.
Three separate augmented VAR models were constructed for three different search terms with the normalized value of the official number of dengue outbreaks data. In each case, an augmented VAR model was constructed by adding an extra lag dmax to the optimal lag length k. For the VAR model between the Google Trends search term “dengue fever” and the normalized value of the number of dengue outbreaks in India, the lag of the search term “dengue fever” at 8 weeks was found to significantly affect the normalized value of the number of dengue outbreaks (p 0.002), while no significant effect was observed with other lags. All the eigenvalues for the model were inside the unit circle. The result suggests that the general interest in the search term “dengue fever” rises 8 weeks before the increase in the dengue outbreak. For the VAR model between the Google Trends search term “dengue symptoms” and the normalized value of the number of dengue outbreaks in India, the lag of the search term “dengue symptoms” at 4 weeks was found to significantly affect the normalized value of the number of dengue outbreaks (p 0.004), while no significant effect was observed with other lags. All the eigenvalues for the model were found within the unit circle. The result suggests that general interest in the search term “dengue symptoms” rises 4 weeks before the increase in the number of dengue outbreaks. For the VAR model between the Google Trends search term “dengue treatment” and the normalized value of the number of dengue outbreaks in India, the lag of the search term “dengue treatment” at 2 weeks (p 0.01) and 8 weeks (p 0.02) was found to have a significant positive effect on the normalized value of the number of dengue outbreaks. On the basis of the eigenvalue of the third VAR model between the Google trend search term “dengue treatment” and the normalized value of the number of dengue outbreaks, the model was found to be unstable. The results of the augmented VAR models are presented in the annexure.
The Granger causality test was then performed using the Toda-Yamamoto test by examining the coefficients of the additional lag terms in the augmented VAR models to assess the causal relationships between each search term and the official dengue outbreak data. The Toda-Yamamoto causality test can be used in the presence of nonstationarity and cointegration. Table 2 presents the results of the Toda-Yamamoto causality test. The null hypothesis that Google Trends data for the terms “dengue fever,” “dengue symptoms,” and “dengue treatment” does not Granger cause the normalized value of the official number of dengue outbreaks in India data was rejected at a 5% significance level with p-values of < 0.001, < 0.001, and < 0.001, respectively.
Table 2. Causality analysis between Google trends data and number dengue outbreak using the Toda-Yamamoto test
Search Term | F-Test | Degrees of Freedom (df1, df2) | p value | Conclusion (Reject H₀) | Direction of causality |
|---|---|---|---|---|---|
Dengue fever | 4.4666 | 10, 44 | < 0.001 | Yes | Google trends data (“Dengue fever”)→Normalized value of number of dengue outbreak |
Dengue symptoms | 4.4043 | 6, 68 | < 0.001 | Yes | Google trends data (“Dengue symptoms”)→Normalized value of number of dengue outbreak |
Dengue treatment | 5.007 | 13, 26 | < 0.001 | Yes | Google trends data (“Dengue treatment”)→Normalized value of number of dengue outbreak |
Discussion
This research explores the use of Google Trends data as a supplementary tool to traditional surveillance methods used for dengue outbreaks in India. It leverages publicly available, reliable data on the number of dengue outbreaks from IDSP and highlights the potential of Google Trends data in predicting and providing early warning signs for dengue outbreaks in India. The findings reveal a strong and positive correlation between the Google Trends data for the searched terms “dengue fever” (r 0.82, p < 0.001), “dengue symptoms” (r 0.80, p < 0.001), and “dengue treatment” (r 0.83, p < 0.001) and the normalized value of the official number of dengue outbreaks in India. A line graph of the 3-week smoothed moving average also demonstrated that the Google Trends data curve peaked before the official outbreak data. Additionally, the causality analysis using the Toda Yamamoto causality test demonstrated significant causality between the three searched terms, Google Trends data, and the normalized value of the official number of dengue outbreaks in India (p < 0.001).
Similar results are reported by a number of studies [5, 14, 15, 16, 17–18]. An analysis in 2018 found a strong positive correlation (r 0.70) with a lag of -2 to -3 weeks between Google Trends data for dengue fever and IDSP data in Haryana, India [5]. Likewise, research in 2019 in Indonesia found a very strong correlation between three different search terms in Google Trends data and the Indonesian national surveillance report for dengue, with a value of r ranging from 0.92 to 0.93. It also showed a linear time series pattern with official data on moving average analysis [14]. There are a few studies where the correlation was found to be lower as compared to our findings [19, 20]. A Google Trend analysis in 2018, Manila, Philippines, found only a moderate positive correlation between Google Trends data and dengue surveillance reports with a time lag of -2 to -3 weeks. Possible reasons for the difference could include variations in Internet penetration, variations in dengue surveillance, and variations in settings. India in the last decade has undergone a digital revolution with large mobile-based internet penetration. The reason could also be due to differences in search terms used or different time periods. The difference could also be due to variation in correlation seen during epidemic and non-epidemic periods. Findings from Venezuela (2004) reported a correlation between Google Trends data for dengue and official reports to vary from 0.85 to 0.65, respectively [15]. Another study in Mexico (2014) found a correlation to be higher between Google Trends data for dengue and official data in states with a high incidence of dengue [16]. Researchers have also found similar correlations between Google Trends data for other acute diseases and their official data [7, 21, 22, 23, 24, 25, 26–27]. A causality analysis in Indonesia in 2019 found significant Granger causality between Google Trends data and dengue outbreaks [28]. Significant Granger causality is also seen between Google Trends data for other diseases such as M-Pox [29]. Google Trends data has been found to predate the surveillance data for diseases like respiratory syncytial virus in Japan and COVID-19 in the United States [30, 31].
The findings from our study illustrate the significant potential of Google Trends as an adjunct tool for dengue outbreak prediction. The search interest for the terms “dengue fever,” “dengue symptoms,” and “dengue treatment” peaked before the official outbreak data, as shown in Figs. 1 and 2, and 3, respectively. On the VAR model as well, the general interest in the search terms “dengue fever” and “dengue symptoms” was found to rise 8 weeks and 4 weeks before the official outbreak data. The Google Trends data search interest provides a lead time, which can be used by the health authorities to prepare for the outbreak, such as awareness campaigns, distribution of nets, and clearing artificial collection of water. The Google trend data can also be integrated with other forecasting systems and social media platforms, such as weather forecasts, to improve the model.
There are certain limitations in our study. Pearson’s correlation analysis was used in the study as part of exploratory analysis, but it depends on the assumption of data stationarity, which is not applicable in our study; therefore, the correlation was complemented with the Toda-Yamamoto causality test. The reliability of Google Trends data can also be influenced by events unrelated to disease occurrence, as the Google Trends data reflect the general interest of the public in the searched terms, which can be influenced by factors unrelated to disease occurrence, such as awareness campaigns, new findings, etc. The Google Trends data also has geographical limitations for areas with poor internet connectivity, such as rural and tribal areas, especially in countries like India, with large populations residing in rural parts. Therefore, for optimal effectiveness, Google Trends data should be integrated with traditional surveillance methods.
Conclusion
In India, dengue is a major public health problem, with a significant burden on the health system. Early intervention has the potential to limit its impact. Our study highlights the potential of utilizing Google Trends data as a supplementary tool to serve as early warning signs for dengue outbreaks in India. It offers a promising avenue for enhancing existing surveillance systems, aiding in more effective and proactive public health responses to dengue outbreaks in the country. Future research should focus on refining methodologies by combining other web-based tools, such as social media trends, along with Google Trends data to enhance predictive capabilities further and address limitations inherent in its use.
Acknowledgements
We would like to acknowledge Mrs. Manisha Rani, Senior data analyst, Cognizant for her valuable assistance in the statistical analysis of the dataset.
Author contributions
G.S., A.J., M.A. wrote the main manuscript. G.S collected and analysed the data. G.S prepared the Figs. 1 and 2, and 3. A.J prepared the Tables 1 and 2. All authors reviewed the final manuscript.
Funding
This study did not receive funding from any organisation.
Data availability
The dataset used in this study is publicly available. The Google Trends data for the search terms can be accessed through the Google Trends website (https://trends.google.com/trends/explore? date=2023-01-01%202023-12-31&geo=IN&q=Dengue%20fever&hl=en-GB). The data on dengue outbreaks in India in the year 2023 can be obtained from the Integrated Disease Surveillance Programme website (https://idsp.mohfw.gov.in/index4.php? lang=1&level=0&linkid=406&lid=3689).
Declarations
Competing interests
The authors declare no competing interests.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1. Shepard, DS; Halasa, YA; Tyagi, BK; Adhish, SV; Nandan, D; Karthiga, KS et al. The inclen study group. Economic and disease burden of dengue illness in India. Am J Trop Med Hyg; 2014; 91,
2. World Health Organization. Dengue and severe dengue [Internet]. Geneva: WHO. 2023 [cited 2024 Oct 3]. Available from: https://www.who.int/news-room/fact-sheets/detail/dengue-and-severe-dengue
3. Ministry of Health and Family Welfare, Government of India. Press release – 2nd August 2024 [Internet]. New Delhi: MoHFW. 2024 [cited 2024 Oct 3]. Available from: https://mohfw.gov.in/?q=/press-info/7479
4. Pilot, E; Nittas, V; Murthy, GVS. The organization, implementation, and functioning of dengue surveillance in India-A systematic scoping review. Int J Environ Res Public Health; 2019; 16,
5. Verma, M; Kishore, K; Kumar, M; Sondh, AR; Aggarwal, G; Kathirvel, S. Google search trends predicting disease outbreaks: an analysis from India. Healthc Inf Res; 2018; 24,
6. Milinovich, GJ; Williams, GM; Clements, AC; Hu, W. Internet-based surveillance systems for monitoring emerging infectious diseases. Lancet Infect Dis; 2014; 14,
7. Kang, M; Zhong, H; He, J; Rutherford, S; Yang, F. Using Google trends for influenza surveillance in South China. PLoS ONE; 2013; 8,
8. Wang, D; Lang, JC; Chen, YH. Assessment of using Google trends for real-time monitoring of infectious disease outbreaks: a measles case study. Sci Rep; 2024; 14,
9. Chan, EH; Sahai, V; Conrad, C; Brownstein, JS. Using web search query data to monitor dengue epidemics: a new model for neglected tropical disease surveillance. PLoS Negl Trop Dis; 2011; 5,
10. Carneiro, HA; Mylonakis, E. Google trends: a web-based tool for real-time surveillance of disease outbreaks. Clin Infect Dis; 2009; 49,
11. Google CA. Google; 2024 [cited 2024 Oct 3]. Available from: https://trends.google.com/trends/
12. Integrated Disease Surveillance Programme (IDSP). Weekly Outbreaks [Internet]. New Delhi: MoHFW, Government of India; 2024 [cited 2024 Oct 3]. Available from: https://idsp.mohfw.gov.in/index4.php?lang=1&level=0&linkid=406&lid=3689
13. R Core Team. R: A language and environment for statistical computing [Internet]. Vienna: R Foundation for Statistical Computing; 2024 [cited 2024 Oct 3]. Available from: https://www.r-project.org/
14. Husnayain, A; Fuad, A; Lazuardi, L. Correlation between Google trends on dengue fever and National surveillance report in Indonesia. Glob Health Action; 2019; 12,
15. Strauss, RA; Castro, JS; Reintjes, R; Torres, JR. Google dengue trends: an indicator of epidemic behavior. The Venezuelan case. Int J Med Inf; 2017; 104, pp. 26-30. [DOI: https://dx.doi.org/10.1016/j.ijmedinf.2017.05.003]
16. Gluskin, RT; Johansson, MA; Santillana, M; Brownstein, JS. Evaluation of Internet-based dengue query data: Google dengue trends. PLoS Negl Trop Dis; 2014; 8,
17. Monnaka, VU; Oliveira, CAC. Google trends correlation and sensitivity for outbreaks of dengue and yellow fever in the state of São Paulo. Einstein (Sao Paulo); 2021; 19, eAO5969. [DOI: https://dx.doi.org/10.31744/einstein_journal/2021AO5969] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34346987]
18. Santangelo, OE; Provenzano, S; Vella, C; Firenze, A; Stacchini, L; Cedrone, F et al. Infodemiology and infoveillance of the four most widespread arbovirus diseases in Italy. Epidemiologia; 2024; 5,
19. Sylvestre, E; Cécilia-Joseph, E; Bouzillé, G; Najioullah, F; Etienne, M; Malouines, F et al. The role of heterogenous Real-world data for dengue surveillance in martinique: observational retrospective study. JMIR Public Health Surveill; 2022; 8,
20. Ho, HT; Carvajal, TM; Bautista, JR; Capistrano, JDR; Viacrusis, KM; Hernandez, LFT et al. Using Google trends to examine the Spatio-Temporal incidence and behavioral patterns of dengue disease: A case study in metropolitan manila, Philippines. Trop Med Infect Dis; 2018; 3,
21. Bhattacharya, S. Predicting emerging and re-emerging disease outbreaks through internet search trends: an analysis from India. AIMS Public Health; 2019; 6,
22. Milinovich GJ, Avril SMR, Clements ACA, Brownstein JS, Tong S, Hu W. Using internet search queries for infectious disease surveillance: screening diseases for suitability. BMC Infect Dis. 2014;14(1). https://doi.org/10.1186/s12879-014-0690-1.
23. Seo, DW; Jo, MW; Sohn, CH; Shin, SY; Lee, J; Yu, M et al. Cumulative query method for influenza surveillance using search engine data. J Med Internet Res; 2014; 16,
24. Cho, S; Sohn, CH; Jo, MW; Shin, SY; Lee, JH; Ryoo, SM et al. Correlation between National influenza surveillance data and Google trends in South Korea. PLoS ONE; 2013; 8,
25. Effenberger, M; Kronbichler, A; Shin, JI; Mayer, G; Tilg, H; Perco, P. Association of the COVID-19 pandemic with internet search volumes: A Google trendstm analysis. Int J Infect Dis; 2020; 95, pp. 192-7.1:CAS:528:DC%2BB3cXosFSqsLg%3D [DOI: https://dx.doi.org/10.1016/j.ijid.2020.04.033] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32305520][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7162745]
26. Wang D, Lang JC, Chen YH. Assessment of using Google Trends for real-time monitoring of infectious disease outbreaks: a measles case study. Scientific Reports [Internet]. 2024 Apr 24 [cited 2024 Sep 30];14(1). Available from: https://www.nature.com/articles/s41598-024-60120-8
27. Usman E, Nindrea RD. The Correlation of Google Trends as an Alternative Information Source in the Early Stages of COVID-19 Outbreak in Indonesia. SRP. 2020;11(9):431–8. https://doi.org/10.31838/srp.2020.9.61
28. Syamsuddin M, Fakhruddin M, Sahetapy-Engel JTM, Soewono E. Causality Analysis of Google Trends and Dengue Incidence in Bandung, Indonesia With Linkage of Digital Data Modeling: Longitudinal Observational Study. J Med Internet Res. 2020;22(7):e17633. https://doi.org/10.2196/17633
29. Yan, W; Du, M; Qin, C; Liu, Q; Wang, Y; Liang, W et al. Association between public attention and Monkeypox epidemic: A global lag-correlation analysis. J Med Virol; 2023; 95,
30. Uda, K; Hagiya, H; Yorifuji, T; Koyama, T; Tsuge, M; Yashiro, M; Tsukahara, H. Correlation between National surveillance and search engine query data on respiratory syncytial virus infections in Japan. BMC Public Health; 2022; 22,
31. Kurian, SJ; Bhatti, AUR; Alvi, MA; Ting, HH; Storlie, C; Wilson, PM et al. Correlations between COVID-19 cases and Google trends data in the united states: A State-by-State analysis. Mayo Clin Proc; 2020; 95,
© The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.