Use of Internet Search Data to Monitor Rotavirus Vaccine Impact in the United States, United Kingdom, and Mexico

Use of Internet Search Data to Monitor Rotavirus Vaccine Impact in the United States, United... Abstract Background Previous studies have found a strong correlation between internet search and public health surveillance data. Less is known about how search data respond to public health interventions, such as vaccination, and the consistency of responses in different countries. In this study, we aimed to study the correlation between internet searches for “rotavirus” and rotavirus disease activity in the United States, United Kingdom, and Mexico before and after introduction of rotavirus vaccine. Methods We compared time series of internet searches for “rotavirus” from Google Trends with rotavirus laboratory reports from the United States and United Kingdom and with hospitalizations for acute gastroenteritis in the United States and Mexico. Using time and location parameters, Google quantifies an internet query share (IQS) to measure the relative search volume for specific terms. We analyzed the correlation between IQS and laboratory and hospitalization data before and after national vaccine introductions. Results There was a strong positive correlation between the rotavirus IQS and laboratory reports in the United States (R2 = 0.79) and United Kingdom (R2 = 0.60) and between the rotavirus IQS and acute gastroenteritis hospitalizations in the United States (R2 = 0.87) and Mexico (R2 = 0.69) (P < .0001 for all correlations). The correlations were stronger in the prevaccine period than in the postvaccine period. After vaccine introduction, the mean rotavirus IQS decreased by 40% (95% confidence interval [CI], 25%–55%) in the United States and by 70% (95% CI, 55%–86%) in Mexico. In the United Kingdom, there was a loss of seasonal variation after vaccine introduction. Conclusions Rotavirus internet search data trends mirrored national rotavirus laboratory trends in the United States and United Kingdom and gastroenteritis-hospitalization data in the United States and Mexico; lower correlations were found after rotavirus vaccine introduction. Internet access is expanding rapidly in countries at all income levels. The World Bank has estimated internet use at 87.4, 91.6, and 44.4 users per 100 people in the United States, United Kingdom, and Mexico, respectively [1]. Many internet users use search engines to identify health information; aggregated search data for monitoring disease activity has shown promise and limitations [2–8]. An example limitation is that users can search for disease terms for reasons other than current illness, and these searches for nonillness reasons might be amplified by transient media coverage and thereby weaken the association between searches and real disease activity. Rotavirus is the most common cause of severe diarrhea in children worldwide [9]. In countries with a national rotavirus vaccination program, disease burden was reduced dramatically after vaccine implementation [10]. In 1 study, Google internet searches for “rotavirus” were correlated strongly with rotavirus disease activity in the United States and United Kingdom from 2004 to 2010 [11]. However, that time period preceded vaccine introduction in the United Kingdom and included only 2 full US postvaccine rotavirus seasons with reasonable vaccine uptake, which limits our ability to make conclusions about the value of search data for monitoring vaccine impact. Furthermore, that study did not include any middle-income countries, where both internet search behavior and disease-surveillance systems might differ from those in high-income countries. In this study, we aimed to assess how well Google internet searches capture rotavirus disease trends as a complement to other approaches for monitoring the effect of rotavirus vaccination programs, especially during a period of rapid change after vaccine implementation in countries with different income levels, vaccine-introduction dates, and vaccine coverage. We compared internet search data with rotavirus laboratory detection and hospitalizations for acute gastroenteritis (AGE) in the United States, United Kingdom, and Mexico, which implemented a national rotavirus vaccination program in 2006, 2013, and 2007, respectively. METHODS We conducted a time-series analysis of internet search and laboratory surveillance data for rotavirus and hospitalization surveillance data for AGE using multiple data sources (detailed later) from the United States, United Kingdom, and Mexico. We measured the correlation between internet search and laboratory or hospitalization data before and after implementation of national rotavirus vaccination programs. Our hypothesis was that disease activity was the main driver of search activity, but we also considered the impact of alternative drivers for internet searches for “rotavirus,” specifically rotavirus vaccine-related news events and norovirus illnesses, which cause a similar clinical syndrome. Internet Search Data Internet search volume was estimated using Google Trends internet query shares (IQSs) (www.google.com/trends), which have been used in other studies on disease monitoring [6]. Google Trends summarizes the relative number of Google web searches for a given search term in a specified time and location. Each data point is normalized to the total number of searches within the time and location range it represents, and the resulting values are scaled to range from 0 to 100. Monthly IQSs for “rotavirus” and “rotavirus vaccine” in the United States, United Kingdom, and Mexico for the time period of January 1, 2004, through September 30, 2015, were downloaded on November 13, 2015. Laboratory Data National laboratory surveillance data for rotavirus cases and norovirus outbreaks were available for the United States and for England and Wales (approximating disease patterns for the United Kingdom). US rotavirus laboratory data from January 2004 to September 2015 were sourced from 371 laboratories that reported rotavirus tests to the National Respiratory and Enteric Virus Surveillance System (NREVSS). A variety of clinical, state, and county laboratories participate in the NREVSS, and laboratory participation has varied over time. For this reason, we used the proportion of stool samples that tested positive for rotavirus per month as an indicator of rotavirus activity. US norovirus outbreak laboratory data were sourced from CaliciNet, a national surveillance network of federal, state, and local public health laboratories. Launched in 2009, CaliciNet collects genetic sequence data on laboratory-confirmed norovirus outbreaks (2 or more norovirus-positive samples) that are reported to 29 state and local health departments across the United States. Rotavirus and norovirus laboratory data for England and Wales from January 2004 to September 2015 were sourced from LabBase2 (which became the Second Generation Surveillance System [SGSS] in December 2014), a well-established laboratory reporting system that routinely collects data from laboratories around the United Kingdom on specimens that test positive for any one of many organisms [12]. Using data from this surveillance network, Public Health England publishes weekly rotavirus and norovirus surveillance reports. Because reporting is thought to be relatively consistent over time but negative results are not reported, we used the number of rotavirus-positive test results for analysis. Hospitalization Data National hospitalization data for AGE of any etiology were available for the United States and Mexico. All-cause AGE trends have been used to assess the impact of rotavirus vaccine on hospitalizations given the high proportion of this syndrome associated with rotavirus and the lack of standardization in testing and coding for rotavirus across hospital sites [13–15]. AGE hospitalizations in the United States were sourced from the State Inpatient Databases (SID) of the Healthcare Cost and Utilization Project (HCUP) (maintained by the Agency for Healthcare Research and Quality), which captures hospitalization data from acute care community hospitals [16]. We restricted our analyses to the 31 states that continually reported data to the SID between January 2004 and December 2013, the most current year for which data were available at the time of this analysis; the population covered by the SID represents 76% of the US children younger than 5 years. HCUP SID data were accessed through an active collaboration between the HCUP and the Centers for Disease Control and Prevention. All-cause AGE hospitalizations, including those with a bacterial, parasitic, viral, or undetermined etiology, were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), codes. Per-capita rates were calculated using a previously described methodology [15] by dividing the monthly number of hospitalizations by the number of children younger than 5 years residing in the participating states. Data on AGE hospitalizations in Mexico from January 2005 (the first year of reliable IQS data for Mexico) to December 2014 were sourced from the National System for Health Informatics, an electronic database for Mexico’s Ministry of Health hospitals. Because the catchment populations of the study hospitals were not known, rates of hospitalization for diarrhea per 10000 all-cause hospitalizations were calculated using the same methodology as that in a previous study of the impact of rotavirus vaccine in Mexico [17]. Prevaccine and Postvaccine Comparisons We evaluated differences in internet searches, rotavirus laboratory reports, and AGE hospitalizations before and after vaccine introduction using unpaired-sample t tests. Boundaries for the prevaccine and postvaccine time periods were defined by the months in which rotavirus vaccine was introduced into the national health system in the United Kingdom (July 2013) and Mexico (May 2007) and by the month in which rotavirus vaccine was recommended by the American Academy of Pediatrics in the United States (January 2007). The month in which the vaccine was introduced was excluded from each analysis. Correlation We calculated the coefficient of determination (R2) to evaluate how closely laboratory and hospitalization data fit the IQS data for all years, and we stratified for the prevaccine and postvaccine time periods for all 3 countries. We also calculated the R2 value from multivariable linear regression models to assess the influence of rotavirus vaccine-related searches and norovirus activity (by including the “rotavirus vaccine” IQS and norovirus laboratory data, respectively, as explanatory variables) in the United States and United Kingdom models. Seasonal Variation In all 3 countries, vaccine introduction has changed the seasonality of rotavirus infections [17–20]. To assess if seasonality in rotavirus IQSs was also affected, we compared the ratio of peak IQS (ie, the month with the highest IQS in each season) to the median IQS in each season. A similar peak-to-average ratio was used previously as an indicator of seasonal intensity of rotavirus disease; a higher ratio reflects greater deviation from average and, thus, greater seasonal variation [21]. RESULTS We found strong positive correlations between rotavirus IQSs and laboratory data for the United States (R2 = 0.78) and United Kingdom (R2 = 0.52) (Figure 1 and Table 1). Similarly, we found strong positive correlations between rotavirus IQSs and AGE hospitalization data for the United States (R2 = 0.86) and Mexico (R2 = 0.69) (Figure 2 and Table 1). In all the analyses, correlations were stronger in the prevaccine period (R2 = 0.83, 0.61, 0.88, and 0.74, respectively) than in the postvaccine period (R2 = 0.68, 0.28, 0.88, and 0.42, respectively) (Table 1). Table 1. Correlation Models between RV IQSs and RV Laboratory Detection or AGE Hospitalization Rates for Children Younger than 5 Years Before and After National RV Vaccination Programs in the United States, United Kingdom, and Mexico Country  Input Variable(s)  RV IQS  All Seasons  Before Vaccine  After Vaccine  R2  P  R2  P  R2  P  US  RV laboratory detection  0.78  <.0001  0.83  <.0001  0.68  <.0001  UK  RV laboratory detection  0.52  <.0001  0.61  <.0001  0.28  .005  US  AGE hospitalization rate for children <5 y old  0.86  <.0001  0.88  <.0001  0.88  <.0001  Mexico  AGE hospitalization rate for children <5 y old  0.69  <.0001  0.74  <.0001  0.42  <.0001  US  RV laboratory detection and RV vaccine IQS  0.82  <.0001, .0042  0.93  <.0001, .0021  0.70  <.0001, .0136  UK  RV laboratory detection and RV vaccine IQS  0.64  <.0001, < .0001  0.67  <.0001, .0008  0.49  .0006, .006  US  RV laboratory detection, RV vaccine IQS, and norovirus outbreaks  NA  NA  0.81  <.0001, .0009, <.0001  UK  RV laboratory detection, RV vaccine IQS, and norovirus laboratory reports  0.66  <.0001, <.0001, .08  0.68  <.0001, .002, .35  0.59  .002, .003, .03  Country  Input Variable(s)  RV IQS  All Seasons  Before Vaccine  After Vaccine  R2  P  R2  P  R2  P  US  RV laboratory detection  0.78  <.0001  0.83  <.0001  0.68  <.0001  UK  RV laboratory detection  0.52  <.0001  0.61  <.0001  0.28  .005  US  AGE hospitalization rate for children <5 y old  0.86  <.0001  0.88  <.0001  0.88  <.0001  Mexico  AGE hospitalization rate for children <5 y old  0.69  <.0001  0.74  <.0001  0.42  <.0001  US  RV laboratory detection and RV vaccine IQS  0.82  <.0001, .0042  0.93  <.0001, .0021  0.70  <.0001, .0136  UK  RV laboratory detection and RV vaccine IQS  0.64  <.0001, < .0001  0.67  <.0001, .0008  0.49  .0006, .006  US  RV laboratory detection, RV vaccine IQS, and norovirus outbreaks  NA  NA  0.81  <.0001, .0009, <.0001  UK  RV laboratory detection, RV vaccine IQS, and norovirus laboratory reports  0.66  <.0001, <.0001, .08  0.68  <.0001, .002, .35  0.59  .002, .003, .03  Abbreviations: AGE, acute gastroenteritis; IQS, internet query share; NA, not applicable; RV, rotavirus; UK, United Kingdom; US, United States. Time periods are as follows: US all seasons, January 2004 to September 2015; US before vaccine, January 2004 to December 2006; US after vaccine, February 2007 to September 2015; UK all seasons, January 2006 to September 2015; UK before vaccine, January 2006 to June 2013; UK after vaccine, August 2013 to September 2015; Mexico all seasons, November 2004 to December 2014; Mexico before vaccine, November 2004 to April 2007; Mexico after vaccine, June 2007 to December 2014; and US AGE hospitalization data, January 2004 to December 2013. UK laboratory data are from England and Wales only. Months of national RV vaccine introduction are excluded. View Large Figure 1. View largeDownload slide Rotavirus internet query shares compared to rotavirus laboratory detection in the United States (A) and United Kingdom (B). UK data are from England and Wales only. Figure 1. View largeDownload slide Rotavirus internet query shares compared to rotavirus laboratory detection in the United States (A) and United Kingdom (B). UK data are from England and Wales only. Figure 2. View largeDownload slide Rotavirus internet query shares compared to acute gastroenteritis (AGE) hospitalization rates for children younger than 5 years in the United States (A) and Mexico (B). US AGE rates are for the 31 states that consistently reported to the State Inpatient Databases from 2004 to 2013. Figure 2. View largeDownload slide Rotavirus internet query shares compared to acute gastroenteritis (AGE) hospitalization rates for children younger than 5 years in the United States (A) and Mexico (B). US AGE rates are for the 31 states that consistently reported to the State Inpatient Databases from 2004 to 2013. In the United States and United Kingdom, months with high norovirus disease activity experienced high rotavirus IQSs despite low laboratory detection of rotavirus (Figure 3). Correlations improved with the additions of rotavirus vaccine IQSs and norovirus disease activity in both the United States and United Kingdom (Table 1). Figure 3. View largeDownload slide Impact of laboratory-confirmed norovirus outbreaks on rotavirus internet query shares (IQS) in the United States (A) and United Kingdom (B), all available data. The y-axis values are Norovirus outbreaks (United States) or laboratory reports (United Kingdom), rotavirus internet query shares, and proportions (United States) or numbers (United Kingdom) of rotavirus-positive laboratory tests. Shaded bars indicate seasons in which peaks in norovirus outbreaks match rotavirus IQS trends despite low rotavirus laboratory activity. UK data are from England and Wales only and exclude July 2013 (vaccine introduction). Figure 3. View largeDownload slide Impact of laboratory-confirmed norovirus outbreaks on rotavirus internet query shares (IQS) in the United States (A) and United Kingdom (B), all available data. The y-axis values are Norovirus outbreaks (United States) or laboratory reports (United Kingdom), rotavirus internet query shares, and proportions (United States) or numbers (United Kingdom) of rotavirus-positive laboratory tests. Shaded bars indicate seasons in which peaks in norovirus outbreaks match rotavirus IQS trends despite low rotavirus laboratory activity. UK data are from England and Wales only and exclude July 2013 (vaccine introduction). The mean monthly US rotavirus IQS decreased from 36.4 to 25.3 after vaccine introduction, a 30.5% (95% confidence interval [CI], 12.9%–48.1%; P = .0008) reduction (Table 2). Similarly, the rotavirus IQS decreased from 32.7 to 10.7 in Mexico after vaccine introduction, a 67.4% (95% CI, 52%–83%; P < .0001) reduction. The rotavirus IQS increased in the United Kingdom after vaccine introduction, but this increase was not statistically significant (7.3% [95% CI, 10.1% decrease to 24.7% increase]; P = .41). Table 2. RV IQSs, RV Laboratory Detection, and AGE Hospitalization Rates for Children Younger than 5 Years Before and After National RV Vaccination Programs in the United States, United Kingdom, and Mexico Country and Input Variables  All Seasons  Before Vaccine  After Vaccine  Decrease After Vaccine (% [95% CI])  P  Mean  95% CI  Mean  95% CI  Mean  95% CI  United States   RV IQS  28.2  25.3–31.1  36.4  27.7–45.2  25.3  22.9–27.7  30.5 (12.9 to 48.1)  .0008   RV-positive laboratory specimens (%)  10.4  8.6–12.1  17.6  12.5–22.6  7.9  6.5–9.3  55.1 (34.1 to 76.1)  <.0001   AGE hospitalization rate for children <5 y old (per 100 000 children)  261.4  231.5–291.2  364.7  287.5–441.8  216.6  194.8–238.3  40.6 (24.3 to 56.9)  <.0001  United Kingdom   RV IQS  13.2  12.2–14.1  12.9  11.8–14.1  13.9  13–14.8  −7.3 (−24.7 to 10.1)  .41   RV-positive laboratory specimens (n)  1097.6  840.1–1355.0  1300.6  981.5–1619.7  395  270.3–519.6  69.6 (23.7 to 115.5)  .0033  Mexico   RV IQS  16.2  13.5–19.0  32.8  24.3–41.4  10.7  9.7–11.7  67.4 (52 to 83)  <.0001   AGE hospitalization rate for children <5 y old (per 10 000 children)  593.9  533.7–654.1  921.5  767.4–1075.6  485.9  440.9–530.9  47.3 (34.7 to 59.8)  <.0001  Country and Input Variables  All Seasons  Before Vaccine  After Vaccine  Decrease After Vaccine (% [95% CI])  P  Mean  95% CI  Mean  95% CI  Mean  95% CI  United States   RV IQS  28.2  25.3–31.1  36.4  27.7–45.2  25.3  22.9–27.7  30.5 (12.9 to 48.1)  .0008   RV-positive laboratory specimens (%)  10.4  8.6–12.1  17.6  12.5–22.6  7.9  6.5–9.3  55.1 (34.1 to 76.1)  <.0001   AGE hospitalization rate for children <5 y old (per 100 000 children)  261.4  231.5–291.2  364.7  287.5–441.8  216.6  194.8–238.3  40.6 (24.3 to 56.9)  <.0001  United Kingdom   RV IQS  13.2  12.2–14.1  12.9  11.8–14.1  13.9  13–14.8  −7.3 (−24.7 to 10.1)  .41   RV-positive laboratory specimens (n)  1097.6  840.1–1355.0  1300.6  981.5–1619.7  395  270.3–519.6  69.6 (23.7 to 115.5)  .0033  Mexico   RV IQS  16.2  13.5–19.0  32.8  24.3–41.4  10.7  9.7–11.7  67.4 (52 to 83)  <.0001   AGE hospitalization rate for children <5 y old (per 10 000 children)  593.9  533.7–654.1  921.5  767.4–1075.6  485.9  440.9–530.9  47.3 (34.7 to 59.8)  <.0001  Abbreviations: AGE, acute gastroenteritis; CI, confidence interval; IQS, internet query share; RV, rotavirus. Time periods are as follows: US all seasons, January 2004 to September 2015; US before vaccine, January 2004 to December 2006; US after vaccine, February 2007 to September 2015; UK all seasons, January 2006 to September 2015; UK before vaccine, January 2006 to June 2013; UK after vaccine, August 2013 to September 2015; Mexico all seasons, November 2004 to December 2014; Mexico before vaccine, November 2004 to April 2007; Mexico after vaccine, June 2007 to December 2014; and US AGE hospitalization data, January 2004 to December 2013. UK laboratory data are from England and Wales only. Months of national RV vaccine introduction are excluded. View Large Seasonal variations in rotavirus IQSs decreased markedly after vaccine introduction in all 3 countries; we found lower peak/median IQS ratios in postvaccine seasons than in the prevaccine seasons (Figure 4). In the United Kingdom, the median IQS increased after vaccine introduction, compared to decreases in the median IQS in the United States and Mexico. Figure 4. View largeDownload slide Seasonal variations in rotavirus internet query shares in the United States (A), United Kingdom (B), and Mexico (C). Seasons are from July of the first year to June of the second year. The months in which rotavirus vaccine was introduced are excluded. Figure 4. View largeDownload slide Seasonal variations in rotavirus internet query shares in the United States (A), United Kingdom (B), and Mexico (C). Seasons are from July of the first year to June of the second year. The months in which rotavirus vaccine was introduced are excluded. CONCLUSIONS Internet searches for “rotavirus,” as measured by Google Trends IQSs, correlated well with laboratory-confirmed rotavirus disease in the United States and United Kingdom and with AGE hospitalizations in the United States and Mexico. Consistent with decreased rotavirus disease activity after the introduction of national rotavirus vaccination programs, there were declines in the IQSs during peak rotavirus seasons in all 3 countries. In each country, lower correlations between rotavirus IQSs and disease activity were found after vaccine introduction. This finding is explained most likely by internet searches for rotavirus motivated by reasons other than illness, such as queries related to vaccination, that are not likely to be seasonal, resulting in relatively elevated IQSs in months of low disease activity. Additional alternative motives for internet searches include rotavirus vaccine-related news events and norovirus outbreaks; adding these inputs improves the postvaccine correlation to nearly match the prevaccine correlation in the United States and United Kingdom. In the United States, the correlation between rotavirus IQSs and AGE hospitalizations was stronger than that between the IQSs and rotavirus laboratory detection The stronger correlation with hospitalizations might be a result of higher specificity for rotavirus disease with laboratory testing than with AGE hospitalizations and rotavirus IQSs. The difference in specificity is evidenced by variability in monthly values that is more similar for rotavirus IQSs and AGE hospitalizations than with laboratory detection. The amplitudes (maximum/minimum value ratio) for rotavirus IQSs and AGE hospitalizations were 9 and 8 (911:115 per 100 000), respectively, whereas for the rotavirus test-positive proportion, the amplitude was 45 (0.45:0.01). The trends in activity were similar for all 3 variables, but the higher variability for the more specific measure of laboratory testing contributed to the lower correlation with the less specific measure of internet searches. We found several outlier monthly IQSs that were likely attributable to news items related to rotavirus vaccines and not disease. On March 22, 2010, the media reported that the US Food and Drug Administration suspended the use of Rotarix after researchers found porcine circovirus genetic material in the vaccines [22]. On November 12, 2012, the United Kingdom announced that rotavirus vaccine would be added to its national program the following summer. Those months (March 2010 for United States and November 2012 for United Kingdom) had extremely high rotavirus IQS scores despite the low level of disease activity, which negatively affected the correlation results. Postvaccine correlations between rotavirus IQSs and rotavirus disease were lowest in the United Kingdom (0.28) and Mexico (0.42). In the United Kingdom, rotavirus IQSs did not decline after vaccine introduction, despite a decline in laboratory detection of rotavirus. This divergence was driven, in part, by searches for rotavirus vaccine, as indicated by the improved correlation when the rotavirus vaccine IQS was added to the model. We also suspect that many vaccine-motivated rotavirus searches did not include “vaccine” in the search term. Although overall searches for rotavirus did not decline, the seasonal intensity of IQSs diminished considerably, consistent with a loss of seasonality in laboratory detection of rotavirus after vaccine introduction. In Mexico, the low postvaccine correlation between rotavirus IQSs and AGE hospitalizations might be a result of alternative etiologies of AGE hospitalization (ie, bacterial enteritis) that are seen more commonly in low- and middle-income countries than in high-income countries [23]. This explanation is supported by the observation that annual peaks in AGE hospitalizations in Mexico changed between the fall-winter seasons before vaccine introduction, when rotavirus was predominant, and the spring-summer seasons after vaccine introduction, when rotavirus was rare [17, 20]. Models for AGE hospitalizations that included IQSs for the bacterial pathogens Shigella and Salmonella (data not shown) resulted in weak individual correlations and did not improve on models with the rotavirus IQS alone. In contrast, the postvaccine correlation between US rotavirus IQSs and AGE hospitalizations remained strong (0.88), because nonrotavirus etiologies of AGE hospitalizations are less common [15, 24, 25]. Our study had several limitations related to how time- series data might not reflect disease activity. Changes in laboratory testing practices might have affected the rotavirus and norovirus laboratory data from the United States and United Kingdom. Changes in nonrotavirus gastroenteritis disease could have affected hospitalization data. Google Trends data could have been affected by changes in who had internet access over time, the type of information available on the internet, and/or the type of information for which people searched during the study time period. Another limitation is generalizability to other countries; differences in internet access according to socioeconomic strata, internet search behavior, and rotavirus testing practices might exist. In conclusion, the results of our study suggest that internet searches can approximate rotavirus disease activity, although results should be contextualized by considering competing search motivations. Internet searches can complement, but not replace, other surveillance approaches for monitoring the impact of rotavirus vaccination programs. Our findings are made more robust by revealing similar results across 3 different countries using both laboratory and epidemiologic measures of disease activity and similar changes after the introduction of rotavirus vaccination programs. Furthermore, divergences in internet search and disease activities can be adjusted for retrospectively in models that consider variables that prompt similar searches, but media events that garner public attention can limit the use of these approaches in real time. Continued worldwide growth in internet access adds promise to the role of big data sources, including internet search, social media, and commercial activities, as a complement to traditional disease surveillance, especially given its low cost, timeliness, and ease of use. Notes Acknowledgments We thank the state partners that provide the SID data to the HCUP. Disclaimer. The findings and conclusions of this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the Agency for Healthcare Research and Quality. Financial support. This work was supported by the National Institute for Health Research Health Protection Research Unit (NIHR HPRU) in gastrointestinal infections at the University of Liverpool in partnership with Public Health England (PHE). The views expressed are those of the authors and not necessarily those of the National Health Service, the NIHR, the Department of Health or Public Health England. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. World Bank. World Development Indicators 2015 . Available at: http://data.worldbank.org/data-catalog/world-development-indicators. Accessed December 17, 2015. 2. Brownstein JS Freifeld CC Madoff LC. Digital disease detection—harnessing the web for public health surveillance. N Engl J Med  2009; 360: 2153– 5. Google Scholar CrossRef Search ADS PubMed  3. Carneiro HA Mylonakis E. Google trends: a web-based tool for real-time surveillance of disease outbreaks. Clin Infect Dis  2009; 49: 1557– 64. Google Scholar CrossRef Search ADS PubMed  4. Ginsberg J Mohebbi MH Patel RSet al.  . Detecting influenza epidemics using search engine query data. Nature  2009; 457: 1012– 4. Google Scholar CrossRef Search ADS PubMed  5. Salathé M Bengtsson L Bodnar TJet al.  . Digital epidemiology. PLoS Comput Biol  2012; 8: e1002616. Google Scholar CrossRef Search ADS PubMed  6. Nuti SV Wayda B Ranasinghe Iet al.  . The use of google trends in health care research: a systematic review. PLoS One  2014; 9: e109583. Google Scholar CrossRef Search ADS PubMed  7. Milinovich GJ Williams GM Clements AC Hu W. Internet-based surveillance systems for monitoring emerging infectious diseases. Lancet Infect Dis  2014; 14: 160– 8. Google Scholar CrossRef Search ADS PubMed  8. Lazer D Kennedy R King G Vespignani A. Big data. The parable of Google flu: traps in big data analysis. Science  2014; 343: 1203– 5. Google Scholar CrossRef Search ADS PubMed  9. Parashar UD Hummelman EG Bresee JSet al.  . Global illness and deaths caused by rotavirus disease in children. Emerg Infect Dis  2003; 9: 565– 72. Google Scholar CrossRef Search ADS PubMed  10. Tate JE Parashar UD. Rotavirus vaccines in routine use. Clin Infect Dis  2014; 59: 1291– 301. Google Scholar CrossRef Search ADS PubMed  11. Desai R Lopman BA Shimshoni Yet al.  . Use of internet search data to monitor impact of rotavirus vaccination in the United States. Clin Infect Dis  2012; 54: e115– 8. Google Scholar CrossRef Search ADS PubMed  12. Wall PG de Louvois J Gilbert RJ Rowe B. Food poisoning: notifications, laboratory reports, and outbreaks—where do the statistics come from and what do they mean? Commun Dis Rep CDR Rev  1996; 6: R93– 100. Google Scholar PubMed  13. Cortes JE Curns AT Tate JEet al.  . Rotavirus vaccine and health care utilization for diarrhea in U.S. children. N Engl J Med  2011; 365: 1108– 17. Google Scholar CrossRef Search ADS PubMed  14. Curns AT Steiner CA Barrett Met al.  . Reduction in acute gastroenteritis hospitalizations among US children after introduction of rotavirus vaccine: analysis of hospital discharge data from 18 US states. J Infect Dis  2010; 201: 1617– 24. Google Scholar CrossRef Search ADS PubMed  15. Leshem E Tate JE Steiner CAet al.  . Acute gastroenteritis hospitalizations among US children following implementation of the rotavirus vaccine. JAMA  2015; 313: 2282– 4. Google Scholar CrossRef Search ADS PubMed  16. HCUP State Inpatient Databases. Overview of the State Inpatient Databases (SID) . Agency for Healthcare Research and Quality, Rockville, MD. Available at: www.hcup-us.ahrq.gov/sidoverview.jsp. Accessed August 5, 2015. 17. Esparza-Aguilar M Gastañaduy PA Sánchez-Uribe Eet al.  . Diarrhoea-related hospitalizations in children before and after implementation of monovalent rotavirus vaccination in Mexico. Bull World Health Organ  2014; 92: 117– 25. Google Scholar CrossRef Search ADS PubMed  18. Tate JE Panozzo CA Payne DCet al.  . Decline and change in seasonality of US rotavirus activity after the introduction of rotavirus vaccine. Pediatrics  2009; 124: 465– 71. Google Scholar CrossRef Search ADS PubMed  19. Atchison C Lopman B Edmunds WJ. Modelling the seasonality of rotavirus disease and the impact of vaccination in England and Wales. Vaccine  2010; 28: 3118– 26. Google Scholar CrossRef Search ADS PubMed  20. Velázquez FR Garcia-Lozano H Rodriguez Eet al.  . Diarrhea morbidity and mortality in Mexican children: impact of rotavirus disease. Pediatr Infect Dis J  2004; 23( 10 Suppl): S149– 55. Google Scholar CrossRef Search ADS PubMed  21. Patel MM Pitzer VE Alonso WJet al.  . Global seasonality of rotavirus disease. Pediatr Infect Dis J  2013; 32: e134– 47. Google Scholar CrossRef Search ADS PubMed  22. Victoria JG Wang C Jones MSet al.  . Viral nucleic acids in live-attenuated vaccines: detection of minority variants and an adventitious virus. J Virol  2010; 84: 6033– 40. Google Scholar CrossRef Search ADS PubMed  23. Kotloff KL Nataro JP Blackwelder WCet al.  . Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study. Lancet  2013; 382: 209– 22. Google Scholar CrossRef Search ADS PubMed  24. Glass RI Kilgore PE Holman RCet al.  . The epidemiology of rotavirus diarrhea in the United States: surveillance and estimates of disease burden. J Infect Dis  1996; 174( Suppl 1): S5– 11. Google Scholar CrossRef Search ADS PubMed  25. Fischer TK Viboud C Parashar Uet al.  . Hospitalizations and deaths from diarrhea and rotavirus among children <5 years of age in the United States, 1993–2003. J Infect Dis  2007; 195: 1117– 25. Google Scholar CrossRef Search ADS PubMed  Published by Oxford University Press on behalf of The Journal of the Pediatric Infectious Diseases Society 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of the Pediatric Infectious Diseases Society Oxford University Press

Use of Internet Search Data to Monitor Rotavirus Vaccine Impact in the United States, United Kingdom, and Mexico

Loading next page...
 
/lp/ou_press/use-of-internet-search-data-to-monitor-rotavirus-vaccine-impact-in-the-B0dOWun2OK
Publisher
Oxford University Press
Copyright
Published by Oxford University Press on behalf of The Journal of the Pediatric Infectious Diseases Society 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US.
ISSN
2048-7193
eISSN
2048-7207
D.O.I.
10.1093/jpids/pix004
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Previous studies have found a strong correlation between internet search and public health surveillance data. Less is known about how search data respond to public health interventions, such as vaccination, and the consistency of responses in different countries. In this study, we aimed to study the correlation between internet searches for “rotavirus” and rotavirus disease activity in the United States, United Kingdom, and Mexico before and after introduction of rotavirus vaccine. Methods We compared time series of internet searches for “rotavirus” from Google Trends with rotavirus laboratory reports from the United States and United Kingdom and with hospitalizations for acute gastroenteritis in the United States and Mexico. Using time and location parameters, Google quantifies an internet query share (IQS) to measure the relative search volume for specific terms. We analyzed the correlation between IQS and laboratory and hospitalization data before and after national vaccine introductions. Results There was a strong positive correlation between the rotavirus IQS and laboratory reports in the United States (R2 = 0.79) and United Kingdom (R2 = 0.60) and between the rotavirus IQS and acute gastroenteritis hospitalizations in the United States (R2 = 0.87) and Mexico (R2 = 0.69) (P < .0001 for all correlations). The correlations were stronger in the prevaccine period than in the postvaccine period. After vaccine introduction, the mean rotavirus IQS decreased by 40% (95% confidence interval [CI], 25%–55%) in the United States and by 70% (95% CI, 55%–86%) in Mexico. In the United Kingdom, there was a loss of seasonal variation after vaccine introduction. Conclusions Rotavirus internet search data trends mirrored national rotavirus laboratory trends in the United States and United Kingdom and gastroenteritis-hospitalization data in the United States and Mexico; lower correlations were found after rotavirus vaccine introduction. Internet access is expanding rapidly in countries at all income levels. The World Bank has estimated internet use at 87.4, 91.6, and 44.4 users per 100 people in the United States, United Kingdom, and Mexico, respectively [1]. Many internet users use search engines to identify health information; aggregated search data for monitoring disease activity has shown promise and limitations [2–8]. An example limitation is that users can search for disease terms for reasons other than current illness, and these searches for nonillness reasons might be amplified by transient media coverage and thereby weaken the association between searches and real disease activity. Rotavirus is the most common cause of severe diarrhea in children worldwide [9]. In countries with a national rotavirus vaccination program, disease burden was reduced dramatically after vaccine implementation [10]. In 1 study, Google internet searches for “rotavirus” were correlated strongly with rotavirus disease activity in the United States and United Kingdom from 2004 to 2010 [11]. However, that time period preceded vaccine introduction in the United Kingdom and included only 2 full US postvaccine rotavirus seasons with reasonable vaccine uptake, which limits our ability to make conclusions about the value of search data for monitoring vaccine impact. Furthermore, that study did not include any middle-income countries, where both internet search behavior and disease-surveillance systems might differ from those in high-income countries. In this study, we aimed to assess how well Google internet searches capture rotavirus disease trends as a complement to other approaches for monitoring the effect of rotavirus vaccination programs, especially during a period of rapid change after vaccine implementation in countries with different income levels, vaccine-introduction dates, and vaccine coverage. We compared internet search data with rotavirus laboratory detection and hospitalizations for acute gastroenteritis (AGE) in the United States, United Kingdom, and Mexico, which implemented a national rotavirus vaccination program in 2006, 2013, and 2007, respectively. METHODS We conducted a time-series analysis of internet search and laboratory surveillance data for rotavirus and hospitalization surveillance data for AGE using multiple data sources (detailed later) from the United States, United Kingdom, and Mexico. We measured the correlation between internet search and laboratory or hospitalization data before and after implementation of national rotavirus vaccination programs. Our hypothesis was that disease activity was the main driver of search activity, but we also considered the impact of alternative drivers for internet searches for “rotavirus,” specifically rotavirus vaccine-related news events and norovirus illnesses, which cause a similar clinical syndrome. Internet Search Data Internet search volume was estimated using Google Trends internet query shares (IQSs) (www.google.com/trends), which have been used in other studies on disease monitoring [6]. Google Trends summarizes the relative number of Google web searches for a given search term in a specified time and location. Each data point is normalized to the total number of searches within the time and location range it represents, and the resulting values are scaled to range from 0 to 100. Monthly IQSs for “rotavirus” and “rotavirus vaccine” in the United States, United Kingdom, and Mexico for the time period of January 1, 2004, through September 30, 2015, were downloaded on November 13, 2015. Laboratory Data National laboratory surveillance data for rotavirus cases and norovirus outbreaks were available for the United States and for England and Wales (approximating disease patterns for the United Kingdom). US rotavirus laboratory data from January 2004 to September 2015 were sourced from 371 laboratories that reported rotavirus tests to the National Respiratory and Enteric Virus Surveillance System (NREVSS). A variety of clinical, state, and county laboratories participate in the NREVSS, and laboratory participation has varied over time. For this reason, we used the proportion of stool samples that tested positive for rotavirus per month as an indicator of rotavirus activity. US norovirus outbreak laboratory data were sourced from CaliciNet, a national surveillance network of federal, state, and local public health laboratories. Launched in 2009, CaliciNet collects genetic sequence data on laboratory-confirmed norovirus outbreaks (2 or more norovirus-positive samples) that are reported to 29 state and local health departments across the United States. Rotavirus and norovirus laboratory data for England and Wales from January 2004 to September 2015 were sourced from LabBase2 (which became the Second Generation Surveillance System [SGSS] in December 2014), a well-established laboratory reporting system that routinely collects data from laboratories around the United Kingdom on specimens that test positive for any one of many organisms [12]. Using data from this surveillance network, Public Health England publishes weekly rotavirus and norovirus surveillance reports. Because reporting is thought to be relatively consistent over time but negative results are not reported, we used the number of rotavirus-positive test results for analysis. Hospitalization Data National hospitalization data for AGE of any etiology were available for the United States and Mexico. All-cause AGE trends have been used to assess the impact of rotavirus vaccine on hospitalizations given the high proportion of this syndrome associated with rotavirus and the lack of standardization in testing and coding for rotavirus across hospital sites [13–15]. AGE hospitalizations in the United States were sourced from the State Inpatient Databases (SID) of the Healthcare Cost and Utilization Project (HCUP) (maintained by the Agency for Healthcare Research and Quality), which captures hospitalization data from acute care community hospitals [16]. We restricted our analyses to the 31 states that continually reported data to the SID between January 2004 and December 2013, the most current year for which data were available at the time of this analysis; the population covered by the SID represents 76% of the US children younger than 5 years. HCUP SID data were accessed through an active collaboration between the HCUP and the Centers for Disease Control and Prevention. All-cause AGE hospitalizations, including those with a bacterial, parasitic, viral, or undetermined etiology, were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), codes. Per-capita rates were calculated using a previously described methodology [15] by dividing the monthly number of hospitalizations by the number of children younger than 5 years residing in the participating states. Data on AGE hospitalizations in Mexico from January 2005 (the first year of reliable IQS data for Mexico) to December 2014 were sourced from the National System for Health Informatics, an electronic database for Mexico’s Ministry of Health hospitals. Because the catchment populations of the study hospitals were not known, rates of hospitalization for diarrhea per 10000 all-cause hospitalizations were calculated using the same methodology as that in a previous study of the impact of rotavirus vaccine in Mexico [17]. Prevaccine and Postvaccine Comparisons We evaluated differences in internet searches, rotavirus laboratory reports, and AGE hospitalizations before and after vaccine introduction using unpaired-sample t tests. Boundaries for the prevaccine and postvaccine time periods were defined by the months in which rotavirus vaccine was introduced into the national health system in the United Kingdom (July 2013) and Mexico (May 2007) and by the month in which rotavirus vaccine was recommended by the American Academy of Pediatrics in the United States (January 2007). The month in which the vaccine was introduced was excluded from each analysis. Correlation We calculated the coefficient of determination (R2) to evaluate how closely laboratory and hospitalization data fit the IQS data for all years, and we stratified for the prevaccine and postvaccine time periods for all 3 countries. We also calculated the R2 value from multivariable linear regression models to assess the influence of rotavirus vaccine-related searches and norovirus activity (by including the “rotavirus vaccine” IQS and norovirus laboratory data, respectively, as explanatory variables) in the United States and United Kingdom models. Seasonal Variation In all 3 countries, vaccine introduction has changed the seasonality of rotavirus infections [17–20]. To assess if seasonality in rotavirus IQSs was also affected, we compared the ratio of peak IQS (ie, the month with the highest IQS in each season) to the median IQS in each season. A similar peak-to-average ratio was used previously as an indicator of seasonal intensity of rotavirus disease; a higher ratio reflects greater deviation from average and, thus, greater seasonal variation [21]. RESULTS We found strong positive correlations between rotavirus IQSs and laboratory data for the United States (R2 = 0.78) and United Kingdom (R2 = 0.52) (Figure 1 and Table 1). Similarly, we found strong positive correlations between rotavirus IQSs and AGE hospitalization data for the United States (R2 = 0.86) and Mexico (R2 = 0.69) (Figure 2 and Table 1). In all the analyses, correlations were stronger in the prevaccine period (R2 = 0.83, 0.61, 0.88, and 0.74, respectively) than in the postvaccine period (R2 = 0.68, 0.28, 0.88, and 0.42, respectively) (Table 1). Table 1. Correlation Models between RV IQSs and RV Laboratory Detection or AGE Hospitalization Rates for Children Younger than 5 Years Before and After National RV Vaccination Programs in the United States, United Kingdom, and Mexico Country  Input Variable(s)  RV IQS  All Seasons  Before Vaccine  After Vaccine  R2  P  R2  P  R2  P  US  RV laboratory detection  0.78  <.0001  0.83  <.0001  0.68  <.0001  UK  RV laboratory detection  0.52  <.0001  0.61  <.0001  0.28  .005  US  AGE hospitalization rate for children <5 y old  0.86  <.0001  0.88  <.0001  0.88  <.0001  Mexico  AGE hospitalization rate for children <5 y old  0.69  <.0001  0.74  <.0001  0.42  <.0001  US  RV laboratory detection and RV vaccine IQS  0.82  <.0001, .0042  0.93  <.0001, .0021  0.70  <.0001, .0136  UK  RV laboratory detection and RV vaccine IQS  0.64  <.0001, < .0001  0.67  <.0001, .0008  0.49  .0006, .006  US  RV laboratory detection, RV vaccine IQS, and norovirus outbreaks  NA  NA  0.81  <.0001, .0009, <.0001  UK  RV laboratory detection, RV vaccine IQS, and norovirus laboratory reports  0.66  <.0001, <.0001, .08  0.68  <.0001, .002, .35  0.59  .002, .003, .03  Country  Input Variable(s)  RV IQS  All Seasons  Before Vaccine  After Vaccine  R2  P  R2  P  R2  P  US  RV laboratory detection  0.78  <.0001  0.83  <.0001  0.68  <.0001  UK  RV laboratory detection  0.52  <.0001  0.61  <.0001  0.28  .005  US  AGE hospitalization rate for children <5 y old  0.86  <.0001  0.88  <.0001  0.88  <.0001  Mexico  AGE hospitalization rate for children <5 y old  0.69  <.0001  0.74  <.0001  0.42  <.0001  US  RV laboratory detection and RV vaccine IQS  0.82  <.0001, .0042  0.93  <.0001, .0021  0.70  <.0001, .0136  UK  RV laboratory detection and RV vaccine IQS  0.64  <.0001, < .0001  0.67  <.0001, .0008  0.49  .0006, .006  US  RV laboratory detection, RV vaccine IQS, and norovirus outbreaks  NA  NA  0.81  <.0001, .0009, <.0001  UK  RV laboratory detection, RV vaccine IQS, and norovirus laboratory reports  0.66  <.0001, <.0001, .08  0.68  <.0001, .002, .35  0.59  .002, .003, .03  Abbreviations: AGE, acute gastroenteritis; IQS, internet query share; NA, not applicable; RV, rotavirus; UK, United Kingdom; US, United States. Time periods are as follows: US all seasons, January 2004 to September 2015; US before vaccine, January 2004 to December 2006; US after vaccine, February 2007 to September 2015; UK all seasons, January 2006 to September 2015; UK before vaccine, January 2006 to June 2013; UK after vaccine, August 2013 to September 2015; Mexico all seasons, November 2004 to December 2014; Mexico before vaccine, November 2004 to April 2007; Mexico after vaccine, June 2007 to December 2014; and US AGE hospitalization data, January 2004 to December 2013. UK laboratory data are from England and Wales only. Months of national RV vaccine introduction are excluded. View Large Figure 1. View largeDownload slide Rotavirus internet query shares compared to rotavirus laboratory detection in the United States (A) and United Kingdom (B). UK data are from England and Wales only. Figure 1. View largeDownload slide Rotavirus internet query shares compared to rotavirus laboratory detection in the United States (A) and United Kingdom (B). UK data are from England and Wales only. Figure 2. View largeDownload slide Rotavirus internet query shares compared to acute gastroenteritis (AGE) hospitalization rates for children younger than 5 years in the United States (A) and Mexico (B). US AGE rates are for the 31 states that consistently reported to the State Inpatient Databases from 2004 to 2013. Figure 2. View largeDownload slide Rotavirus internet query shares compared to acute gastroenteritis (AGE) hospitalization rates for children younger than 5 years in the United States (A) and Mexico (B). US AGE rates are for the 31 states that consistently reported to the State Inpatient Databases from 2004 to 2013. In the United States and United Kingdom, months with high norovirus disease activity experienced high rotavirus IQSs despite low laboratory detection of rotavirus (Figure 3). Correlations improved with the additions of rotavirus vaccine IQSs and norovirus disease activity in both the United States and United Kingdom (Table 1). Figure 3. View largeDownload slide Impact of laboratory-confirmed norovirus outbreaks on rotavirus internet query shares (IQS) in the United States (A) and United Kingdom (B), all available data. The y-axis values are Norovirus outbreaks (United States) or laboratory reports (United Kingdom), rotavirus internet query shares, and proportions (United States) or numbers (United Kingdom) of rotavirus-positive laboratory tests. Shaded bars indicate seasons in which peaks in norovirus outbreaks match rotavirus IQS trends despite low rotavirus laboratory activity. UK data are from England and Wales only and exclude July 2013 (vaccine introduction). Figure 3. View largeDownload slide Impact of laboratory-confirmed norovirus outbreaks on rotavirus internet query shares (IQS) in the United States (A) and United Kingdom (B), all available data. The y-axis values are Norovirus outbreaks (United States) or laboratory reports (United Kingdom), rotavirus internet query shares, and proportions (United States) or numbers (United Kingdom) of rotavirus-positive laboratory tests. Shaded bars indicate seasons in which peaks in norovirus outbreaks match rotavirus IQS trends despite low rotavirus laboratory activity. UK data are from England and Wales only and exclude July 2013 (vaccine introduction). The mean monthly US rotavirus IQS decreased from 36.4 to 25.3 after vaccine introduction, a 30.5% (95% confidence interval [CI], 12.9%–48.1%; P = .0008) reduction (Table 2). Similarly, the rotavirus IQS decreased from 32.7 to 10.7 in Mexico after vaccine introduction, a 67.4% (95% CI, 52%–83%; P < .0001) reduction. The rotavirus IQS increased in the United Kingdom after vaccine introduction, but this increase was not statistically significant (7.3% [95% CI, 10.1% decrease to 24.7% increase]; P = .41). Table 2. RV IQSs, RV Laboratory Detection, and AGE Hospitalization Rates for Children Younger than 5 Years Before and After National RV Vaccination Programs in the United States, United Kingdom, and Mexico Country and Input Variables  All Seasons  Before Vaccine  After Vaccine  Decrease After Vaccine (% [95% CI])  P  Mean  95% CI  Mean  95% CI  Mean  95% CI  United States   RV IQS  28.2  25.3–31.1  36.4  27.7–45.2  25.3  22.9–27.7  30.5 (12.9 to 48.1)  .0008   RV-positive laboratory specimens (%)  10.4  8.6–12.1  17.6  12.5–22.6  7.9  6.5–9.3  55.1 (34.1 to 76.1)  <.0001   AGE hospitalization rate for children <5 y old (per 100 000 children)  261.4  231.5–291.2  364.7  287.5–441.8  216.6  194.8–238.3  40.6 (24.3 to 56.9)  <.0001  United Kingdom   RV IQS  13.2  12.2–14.1  12.9  11.8–14.1  13.9  13–14.8  −7.3 (−24.7 to 10.1)  .41   RV-positive laboratory specimens (n)  1097.6  840.1–1355.0  1300.6  981.5–1619.7  395  270.3–519.6  69.6 (23.7 to 115.5)  .0033  Mexico   RV IQS  16.2  13.5–19.0  32.8  24.3–41.4  10.7  9.7–11.7  67.4 (52 to 83)  <.0001   AGE hospitalization rate for children <5 y old (per 10 000 children)  593.9  533.7–654.1  921.5  767.4–1075.6  485.9  440.9–530.9  47.3 (34.7 to 59.8)  <.0001  Country and Input Variables  All Seasons  Before Vaccine  After Vaccine  Decrease After Vaccine (% [95% CI])  P  Mean  95% CI  Mean  95% CI  Mean  95% CI  United States   RV IQS  28.2  25.3–31.1  36.4  27.7–45.2  25.3  22.9–27.7  30.5 (12.9 to 48.1)  .0008   RV-positive laboratory specimens (%)  10.4  8.6–12.1  17.6  12.5–22.6  7.9  6.5–9.3  55.1 (34.1 to 76.1)  <.0001   AGE hospitalization rate for children <5 y old (per 100 000 children)  261.4  231.5–291.2  364.7  287.5–441.8  216.6  194.8–238.3  40.6 (24.3 to 56.9)  <.0001  United Kingdom   RV IQS  13.2  12.2–14.1  12.9  11.8–14.1  13.9  13–14.8  −7.3 (−24.7 to 10.1)  .41   RV-positive laboratory specimens (n)  1097.6  840.1–1355.0  1300.6  981.5–1619.7  395  270.3–519.6  69.6 (23.7 to 115.5)  .0033  Mexico   RV IQS  16.2  13.5–19.0  32.8  24.3–41.4  10.7  9.7–11.7  67.4 (52 to 83)  <.0001   AGE hospitalization rate for children <5 y old (per 10 000 children)  593.9  533.7–654.1  921.5  767.4–1075.6  485.9  440.9–530.9  47.3 (34.7 to 59.8)  <.0001  Abbreviations: AGE, acute gastroenteritis; CI, confidence interval; IQS, internet query share; RV, rotavirus. Time periods are as follows: US all seasons, January 2004 to September 2015; US before vaccine, January 2004 to December 2006; US after vaccine, February 2007 to September 2015; UK all seasons, January 2006 to September 2015; UK before vaccine, January 2006 to June 2013; UK after vaccine, August 2013 to September 2015; Mexico all seasons, November 2004 to December 2014; Mexico before vaccine, November 2004 to April 2007; Mexico after vaccine, June 2007 to December 2014; and US AGE hospitalization data, January 2004 to December 2013. UK laboratory data are from England and Wales only. Months of national RV vaccine introduction are excluded. View Large Seasonal variations in rotavirus IQSs decreased markedly after vaccine introduction in all 3 countries; we found lower peak/median IQS ratios in postvaccine seasons than in the prevaccine seasons (Figure 4). In the United Kingdom, the median IQS increased after vaccine introduction, compared to decreases in the median IQS in the United States and Mexico. Figure 4. View largeDownload slide Seasonal variations in rotavirus internet query shares in the United States (A), United Kingdom (B), and Mexico (C). Seasons are from July of the first year to June of the second year. The months in which rotavirus vaccine was introduced are excluded. Figure 4. View largeDownload slide Seasonal variations in rotavirus internet query shares in the United States (A), United Kingdom (B), and Mexico (C). Seasons are from July of the first year to June of the second year. The months in which rotavirus vaccine was introduced are excluded. CONCLUSIONS Internet searches for “rotavirus,” as measured by Google Trends IQSs, correlated well with laboratory-confirmed rotavirus disease in the United States and United Kingdom and with AGE hospitalizations in the United States and Mexico. Consistent with decreased rotavirus disease activity after the introduction of national rotavirus vaccination programs, there were declines in the IQSs during peak rotavirus seasons in all 3 countries. In each country, lower correlations between rotavirus IQSs and disease activity were found after vaccine introduction. This finding is explained most likely by internet searches for rotavirus motivated by reasons other than illness, such as queries related to vaccination, that are not likely to be seasonal, resulting in relatively elevated IQSs in months of low disease activity. Additional alternative motives for internet searches include rotavirus vaccine-related news events and norovirus outbreaks; adding these inputs improves the postvaccine correlation to nearly match the prevaccine correlation in the United States and United Kingdom. In the United States, the correlation between rotavirus IQSs and AGE hospitalizations was stronger than that between the IQSs and rotavirus laboratory detection The stronger correlation with hospitalizations might be a result of higher specificity for rotavirus disease with laboratory testing than with AGE hospitalizations and rotavirus IQSs. The difference in specificity is evidenced by variability in monthly values that is more similar for rotavirus IQSs and AGE hospitalizations than with laboratory detection. The amplitudes (maximum/minimum value ratio) for rotavirus IQSs and AGE hospitalizations were 9 and 8 (911:115 per 100 000), respectively, whereas for the rotavirus test-positive proportion, the amplitude was 45 (0.45:0.01). The trends in activity were similar for all 3 variables, but the higher variability for the more specific measure of laboratory testing contributed to the lower correlation with the less specific measure of internet searches. We found several outlier monthly IQSs that were likely attributable to news items related to rotavirus vaccines and not disease. On March 22, 2010, the media reported that the US Food and Drug Administration suspended the use of Rotarix after researchers found porcine circovirus genetic material in the vaccines [22]. On November 12, 2012, the United Kingdom announced that rotavirus vaccine would be added to its national program the following summer. Those months (March 2010 for United States and November 2012 for United Kingdom) had extremely high rotavirus IQS scores despite the low level of disease activity, which negatively affected the correlation results. Postvaccine correlations between rotavirus IQSs and rotavirus disease were lowest in the United Kingdom (0.28) and Mexico (0.42). In the United Kingdom, rotavirus IQSs did not decline after vaccine introduction, despite a decline in laboratory detection of rotavirus. This divergence was driven, in part, by searches for rotavirus vaccine, as indicated by the improved correlation when the rotavirus vaccine IQS was added to the model. We also suspect that many vaccine-motivated rotavirus searches did not include “vaccine” in the search term. Although overall searches for rotavirus did not decline, the seasonal intensity of IQSs diminished considerably, consistent with a loss of seasonality in laboratory detection of rotavirus after vaccine introduction. In Mexico, the low postvaccine correlation between rotavirus IQSs and AGE hospitalizations might be a result of alternative etiologies of AGE hospitalization (ie, bacterial enteritis) that are seen more commonly in low- and middle-income countries than in high-income countries [23]. This explanation is supported by the observation that annual peaks in AGE hospitalizations in Mexico changed between the fall-winter seasons before vaccine introduction, when rotavirus was predominant, and the spring-summer seasons after vaccine introduction, when rotavirus was rare [17, 20]. Models for AGE hospitalizations that included IQSs for the bacterial pathogens Shigella and Salmonella (data not shown) resulted in weak individual correlations and did not improve on models with the rotavirus IQS alone. In contrast, the postvaccine correlation between US rotavirus IQSs and AGE hospitalizations remained strong (0.88), because nonrotavirus etiologies of AGE hospitalizations are less common [15, 24, 25]. Our study had several limitations related to how time- series data might not reflect disease activity. Changes in laboratory testing practices might have affected the rotavirus and norovirus laboratory data from the United States and United Kingdom. Changes in nonrotavirus gastroenteritis disease could have affected hospitalization data. Google Trends data could have been affected by changes in who had internet access over time, the type of information available on the internet, and/or the type of information for which people searched during the study time period. Another limitation is generalizability to other countries; differences in internet access according to socioeconomic strata, internet search behavior, and rotavirus testing practices might exist. In conclusion, the results of our study suggest that internet searches can approximate rotavirus disease activity, although results should be contextualized by considering competing search motivations. Internet searches can complement, but not replace, other surveillance approaches for monitoring the impact of rotavirus vaccination programs. Our findings are made more robust by revealing similar results across 3 different countries using both laboratory and epidemiologic measures of disease activity and similar changes after the introduction of rotavirus vaccination programs. Furthermore, divergences in internet search and disease activities can be adjusted for retrospectively in models that consider variables that prompt similar searches, but media events that garner public attention can limit the use of these approaches in real time. Continued worldwide growth in internet access adds promise to the role of big data sources, including internet search, social media, and commercial activities, as a complement to traditional disease surveillance, especially given its low cost, timeliness, and ease of use. Notes Acknowledgments We thank the state partners that provide the SID data to the HCUP. Disclaimer. The findings and conclusions of this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the Agency for Healthcare Research and Quality. Financial support. This work was supported by the National Institute for Health Research Health Protection Research Unit (NIHR HPRU) in gastrointestinal infections at the University of Liverpool in partnership with Public Health England (PHE). The views expressed are those of the authors and not necessarily those of the National Health Service, the NIHR, the Department of Health or Public Health England. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. World Bank. World Development Indicators 2015 . Available at: http://data.worldbank.org/data-catalog/world-development-indicators. Accessed December 17, 2015. 2. Brownstein JS Freifeld CC Madoff LC. Digital disease detection—harnessing the web for public health surveillance. N Engl J Med  2009; 360: 2153– 5. Google Scholar CrossRef Search ADS PubMed  3. Carneiro HA Mylonakis E. Google trends: a web-based tool for real-time surveillance of disease outbreaks. Clin Infect Dis  2009; 49: 1557– 64. Google Scholar CrossRef Search ADS PubMed  4. Ginsberg J Mohebbi MH Patel RSet al.  . Detecting influenza epidemics using search engine query data. Nature  2009; 457: 1012– 4. Google Scholar CrossRef Search ADS PubMed  5. Salathé M Bengtsson L Bodnar TJet al.  . Digital epidemiology. PLoS Comput Biol  2012; 8: e1002616. Google Scholar CrossRef Search ADS PubMed  6. Nuti SV Wayda B Ranasinghe Iet al.  . The use of google trends in health care research: a systematic review. PLoS One  2014; 9: e109583. Google Scholar CrossRef Search ADS PubMed  7. Milinovich GJ Williams GM Clements AC Hu W. Internet-based surveillance systems for monitoring emerging infectious diseases. Lancet Infect Dis  2014; 14: 160– 8. Google Scholar CrossRef Search ADS PubMed  8. Lazer D Kennedy R King G Vespignani A. Big data. The parable of Google flu: traps in big data analysis. Science  2014; 343: 1203– 5. Google Scholar CrossRef Search ADS PubMed  9. Parashar UD Hummelman EG Bresee JSet al.  . Global illness and deaths caused by rotavirus disease in children. Emerg Infect Dis  2003; 9: 565– 72. Google Scholar CrossRef Search ADS PubMed  10. Tate JE Parashar UD. Rotavirus vaccines in routine use. Clin Infect Dis  2014; 59: 1291– 301. Google Scholar CrossRef Search ADS PubMed  11. Desai R Lopman BA Shimshoni Yet al.  . Use of internet search data to monitor impact of rotavirus vaccination in the United States. Clin Infect Dis  2012; 54: e115– 8. Google Scholar CrossRef Search ADS PubMed  12. Wall PG de Louvois J Gilbert RJ Rowe B. Food poisoning: notifications, laboratory reports, and outbreaks—where do the statistics come from and what do they mean? Commun Dis Rep CDR Rev  1996; 6: R93– 100. Google Scholar PubMed  13. Cortes JE Curns AT Tate JEet al.  . Rotavirus vaccine and health care utilization for diarrhea in U.S. children. N Engl J Med  2011; 365: 1108– 17. Google Scholar CrossRef Search ADS PubMed  14. Curns AT Steiner CA Barrett Met al.  . Reduction in acute gastroenteritis hospitalizations among US children after introduction of rotavirus vaccine: analysis of hospital discharge data from 18 US states. J Infect Dis  2010; 201: 1617– 24. Google Scholar CrossRef Search ADS PubMed  15. Leshem E Tate JE Steiner CAet al.  . Acute gastroenteritis hospitalizations among US children following implementation of the rotavirus vaccine. JAMA  2015; 313: 2282– 4. Google Scholar CrossRef Search ADS PubMed  16. HCUP State Inpatient Databases. Overview of the State Inpatient Databases (SID) . Agency for Healthcare Research and Quality, Rockville, MD. Available at: www.hcup-us.ahrq.gov/sidoverview.jsp. Accessed August 5, 2015. 17. Esparza-Aguilar M Gastañaduy PA Sánchez-Uribe Eet al.  . Diarrhoea-related hospitalizations in children before and after implementation of monovalent rotavirus vaccination in Mexico. Bull World Health Organ  2014; 92: 117– 25. Google Scholar CrossRef Search ADS PubMed  18. Tate JE Panozzo CA Payne DCet al.  . Decline and change in seasonality of US rotavirus activity after the introduction of rotavirus vaccine. Pediatrics  2009; 124: 465– 71. Google Scholar CrossRef Search ADS PubMed  19. Atchison C Lopman B Edmunds WJ. Modelling the seasonality of rotavirus disease and the impact of vaccination in England and Wales. Vaccine  2010; 28: 3118– 26. Google Scholar CrossRef Search ADS PubMed  20. Velázquez FR Garcia-Lozano H Rodriguez Eet al.  . Diarrhea morbidity and mortality in Mexican children: impact of rotavirus disease. Pediatr Infect Dis J  2004; 23( 10 Suppl): S149– 55. Google Scholar CrossRef Search ADS PubMed  21. Patel MM Pitzer VE Alonso WJet al.  . Global seasonality of rotavirus disease. Pediatr Infect Dis J  2013; 32: e134– 47. Google Scholar CrossRef Search ADS PubMed  22. Victoria JG Wang C Jones MSet al.  . Viral nucleic acids in live-attenuated vaccines: detection of minority variants and an adventitious virus. J Virol  2010; 84: 6033– 40. Google Scholar CrossRef Search ADS PubMed  23. Kotloff KL Nataro JP Blackwelder WCet al.  . Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study. Lancet  2013; 382: 209– 22. Google Scholar CrossRef Search ADS PubMed  24. Glass RI Kilgore PE Holman RCet al.  . The epidemiology of rotavirus diarrhea in the United States: surveillance and estimates of disease burden. J Infect Dis  1996; 174( Suppl 1): S5– 11. Google Scholar CrossRef Search ADS PubMed  25. Fischer TK Viboud C Parashar Uet al.  . Hospitalizations and deaths from diarrhea and rotavirus among children <5 years of age in the United States, 1993–2003. J Infect Dis  2007; 195: 1117– 25. Google Scholar CrossRef Search ADS PubMed  Published by Oxford University Press on behalf of The Journal of the Pediatric Infectious Diseases Society 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Journal

Journal of the Pediatric Infectious Diseases SocietyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off