The correlation between Google trends and salmonellosis

Ming-Yang Wang; Nai-Jun Tang

doi:10.1186/s12889-021-11615-w

The correlation between Google trends and salmonellosis

BMC Public Health. 2021 Aug 21;21(1):1575. doi: 10.1186/s12889-021-11615-w.

Authors

Ming-Yang Wang^{1

2

3}, Nai-Jun Tang^{4

5}

Affiliations

¹ Department of Occupational and Environmental Health, School of Public Health, Tianjin Medical University, Tianjin, 300070, China.
² Tianjin Key Laboratory of Environment, Nutrition, and Public Health, Tianjin Medical University, Tianjin, 300070, China.
³ Beijing Tongren Eye Center, Beijing Ophthalmology& Visual Sciences Key Laboratory, Beijing Tongren Hospital Affiliated to Capital University of Medical Sciences, Beijing, 100730, China.
⁴ Department of Occupational and Environmental Health, School of Public Health, Tianjin Medical University, Tianjin, 300070, China. tangnaijun@tmu.edu.cn.
⁵ Tianjin Key Laboratory of Environment, Nutrition, and Public Health, Tianjin Medical University, Tianjin, 300070, China. tangnaijun@tmu.edu.cn.

Abstract

Background: Salmonella infection (salmonellosis) is a common infectious disease leading to gastroenteritis, dehydration, uveitis, etc. Internet search is a new method to monitor the outbreak of infectious disease. An internet-based surveillance system using internet data is logistically advantageous and economical to show term-related diseases. In this study, we tried to determine the relationship between salmonellosis and Google Trends in the USA from January 2004 to December 2017.

Methods: We downloaded the reported salmonellosis in the USA from the National Outbreak Reporting System (NORS) from January 2004 to December 2017. Additionally, we downloaded the Google search terms related to salmonellosis from Google Trends in the same period. Cross-correlation analysis and multiple regression analysis were conducted.

Results: The results showed that 6 Google Trends search terms appeared earlier than reported salmonellosis, 26 Google Trends search terms coincided with salmonellosis, and 16 Google Trends search terms appeared after salmonellosis were reported. When the search terms preceded outbreaks, "foods" (t = 2.927, P = 0.004) was a predictor of salmonellosis. When the search terms coincided with outbreaks, "hotel" (t = 1.854, P = 0.066), "poor sanitation" (t = 2.895, P = 0.004), "blueberries" (t = 2.441, P = 0.016), and "hypovolemic shock" (t = 2.001, P = 0.047) were predictors of salmonellosis. When the search terms appeared after outbreaks, "ice cream" (t = 3.077, P = 0.002) was the predictor of salmonellosis. Finally, we identified the most important indicators of Google Trends search terms, including "hotel" (t = 1.854, P = 0.066), "poor sanitation" (t = 2.895, P = 0.004), "blueberries" (t = 2.441, P = 0.016), and "hypovolemic shock" (t = 2.001, P = 0.047). In the future, the increased search activities of these terms might indicate the salmonellosis.

Conclusion: We evaluated the related Google Trends search terms with salmonellosis and identified the most important predictors of salmonellosis outbreak.

Keywords: Correlation; Google trends; Predictor; Salmonellosis; salmonella.

MeSH terms

Disease Outbreaks
Forecasting
Gastroenteritis* / epidemiology
Humans
Internet
Salmonella Infections*
Search Engine