Temporal Relationship between Healthcare-Associated and Non-Healthcare-Associated Norovirus Outbreaks, and Google Trend in the United States Pubblico
Osuka, Hanako (2017)
Published
Abstract
Background
Norovirus is the leading cause of acute gastroenteritis in the United States. Because it is highly contagious, controlling norovirus outbreaks in healthcare settings is challenging. Since seasonal norovirus activity varies from year to year and rapid implementation of contact precautions is essential, early detection and prediction of the norovirus season may be useful for infection control in healthcare settings. Digital data, such as Google Trends and Twitter, have been increasingly used as a data source to examine infectious disease dynamics. We examined temporal relationships of norovirus outbreaks among healthcare settings, non-healthcare settings, and Google Trends search activity.
Methods
We analyzed norovirus outbreaks from 2009 - 2015 obtained from the National Outbreak Reporting System (NORS) database and Google Trends data in the same period in the United States. We categorized outbreaks into healthcare-associated and non-healthcare-associated, then examined temporal relationships between healthcare-associated outbreaks with (a) non-healthcare-associated norovirus outbreaks and (b) Google Trends data. We identified and compared the onset, peak, and end of the season and conducted linear regression analysis with a series of lags.
Results
11,212 confirmed and suspected norovirus outbreaks involving a total of 397,148 primary cases were reported to NORS during 2009-2015. Healthcare-associated outbreaks had more pronounced seasonality than non-healthcare outbreaks, as they had a higher peak-mean ratio (5.5 v.s. 3.3) and were more concentrated in winter; 63.6% v.s. 44.6% of total outbreaks occurred during November to February. There was weak correlations between weekly counts of healthcare-associated outbreaks with (a) non-healthcare-associated outbreaks (R2 = 0.39) and (b) moderate correlation with Google Trends activity (R2 = 0.68) overall. During the increasing phase of the season, healthcare-associated and non-healthcare-associated outbreaks with a seven-week lead showed the highest correlation (R2 = 0.43). The strongest correlation was observed with no time lag between Google Trends activity and healthcare-associated outbreaks during the increasing phase of the season (R2 = 0.68).
Conclusions
Non-healthcare-associated norovirus outbreaks are less seasonal but increased earlier than healthcare-associated outbreaks. Google Trends data showed moderate correlation with healthcare-associated outbreaks, but without preceding lag. Monitoring community norovirus activity and Google Trends data may have a potential to supplement existing norovirus surveillance and provide early warning of the season.
Table of Contents
BACKGROUNDS…………………………………………………………………………… 1
METHODS ………………………………………………………………………………….. 5
RESULTS …………………………………………………………………………………… 10
DISCUSSION ……………………………………………………………………………… 14
FIGURES …………………………………………………………………………………… 20
Figure 1. Number of Weekly Reported Healthcare-Associated and Non-Healthcare-Associated Norovirus Outbreaks and Google Trends Score with "Stomach Virus" between 2009 and 2015, United States
Figure 2. The Week of Onset, Peak, and End of Season for Healthcare-Associated and Non-Healthcare-Associated Norovirus Outbreaks and Google Trends Score with "Stomach Virus" between 2009 and 2015, United States
TABLES ……………………………………………………………………………………... 22
Table 1. Number and Percentage of Reported Norovirus Outbreaks by Setting, National Outbreak Reporting System, 2009-2015, United States
Table 2. Pearson Correlation Coefficients Matrix among Outbreaks Settings, National Outbreak Reporting System, 2009-2015, United States
Table 3. Characteristics of Outbreak by Each Setting, from the National Outbreak Reporting System (NORS) Database, 2009-2015, the United States
Table 4. R2 between Monthly Number of Hospital-Associated Norovirus Outbreak and Monthly Google Trends Activity
Table 5. The Coefficient with 95% Confidence Interval, R-square with the Regression Model with Lag-time
Table 6. The Coefficient with 95% Confidence Interval, R-square with the Regression Model with Lag-time: Before and After the Peak
REFERENCES ……………………………………………………………………………… 29
About this Master's Thesis
| School | |
|---|---|
| Department | |
| Degree | |
| Submission | |
| Language |
|
| Research Field | |
| Parola chiave | |
| Committee Chair / Thesis Advisor | |
| Committee Members | |
| Partnering Agencies |
Primary PDF
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
|
Temporal Relationship between Healthcare-Associated and Non-Healthcare-Associated Norovirus Outbreaks, and Google Trend in the United States () | 2018-08-28 16:02:55 -0400 |
|
Supplemental Files
| Thumbnail | Title | Date Uploaded | Actions |
|---|