key: cord-0889162-lyejc0f7
authors: Cook, Alex R.; Chen, Mark I.C.; Lin, Raymond Tzer Pin
title: Internet Search Limitations and Pandemic Influenza, Singapore
date: 2010-10-03
journal: Emerg Infect Dis
DOI: 10.3201/eid1610.100840
sha: ff7d49ac4008f60ef9c5a437e0d504dcefd1246f
doc_id: 889162
cord_uid: lyejc0f7

nan

results of studies conducted in other countries (5) (6) (7) . None of the pregnant women were in critical condition and none died of pandemic (H1N1) 2009 infection during the surveillance period. The overall lower prevalence of pregnant women in Japan (≈0.67% of the total population) compared with that in other countries (8) might account for the low number of pregnant women hospitalized for pandemic (H1N1) 2009 virus in Japan.

Of all 12,702 patients, 4,420 (34.8%) had underlying medical conditions. In terms of age, 1,848 (32.6%) of 5,670 patients 5-9 years of age and 324 (81.0%) of 400 patients >70 years of age had underlying medical conditions. For those with underlying medical conditions, incidence of hospitalization was relatively higher for children than for adults; rates for hospitalized children were 814.1, 573.7, and 219.0 per 100,000 persons for those 5-9, 10-14, and 15-19 years of age, respectively; whereas, the rate for hospitalized adults 60-69 years of age was 6.7/100,000 persons. In all age categories, the incidence of hospitalization for pandemic (H1N1) 2009 virus infection was notably higher for patients with underlying medical conditions than for those without. The attributable risk for medical conditions associated with hospitalization for pandemic (H1N1) 2009 virus infection was highest among patients 5-9 and 10-14 years of age (744.7 and 549.6, respectively), whereas the risk was lowest (6.5) among patients 60-69 years of age. The relative risk associated with underlying medical conditions was comparatively higher for patients in age groups 15-19, 20-29, and 30-39 years (48.4, 54.9, and 48.4, respectively) than among those >70 years of age (17.9). In contrast to seasonal infl uenza (4), persons <20 years of age had higher risk for hospitalization for pandemic (H1N1) 2009 than did older age groups.

Considering our fi ndings regarding the attributable risk for hospitalization, interventions that aim to control pandemic (H1N1) 2009 virus infection among children, especially those with underlying medical conditions, can be considered key for minimizing the strain (fi nancial, staffi ng, space) on the healthcare system. Our fi ndings justify prioritizing the treatment of children and young adults by vaccination and early prescription of antiviral drugs. 

To the Editor: In the past few years, several publications have reported that Internet search queries may usefully supplement other, traditional surveillance programs for infectious diseases (1-3). The philanthropic arm of Google offers Flu Trends, a site that provides up-to-date estimates of infl uenza activity in 20 countries of the Pacifi c Rim and Europe (4) by using data mining techniques to fi nd good predictors of historic infl uenza indicators (1).

This service has yet to be extended to other countries and other diseases because access to offi cial surveillance data is required, among other reasons. However, another Google service, Insights for Search, enables users to fi nd and download time-series data of relative counts of arbitrary searches for a large number of countries (5). Pelat et al. have shown that a few, well-chosen searches on Google Insights provide data that closely correlate with French surveillance data for seasonal infl uenza, chickenpox, and gastroenteritis (3) . Although Internet searches appear to be a promising tool for public health surveillance, our experience from using Google Insights in the context of pandemic (H1N1) 2009 in Singapore suggests it has important limitations.

In Singapore, the recent pandemic caused an outbreak that peaked at the start of August 2009; the fi rst confi rmed importation was at the end of May and fi rst confi rmed unlinked case was at the end of June. However, the number Google searches for "infl uenza," "H1N1," "swine fl u," and similar terms (in English and Chinese), as well as symptoms associated with the disease, peaked much earlier than did the number of cases (Figure) . The number of searches surged after newsworthy events but was low during the epidemic itself and had declined to about 20% of maximum search volume by the time of the actual peak, as shown by traditional surveillance. Fur-thermore, no discernible local maxima were observed that corresponded to the peak in case data. In contrast, alternative traditional measures of infl uenza incidence-prevalence of the novel strain among viral samples and general practice surveillance (6,7)-provide a consistent description of the outbreak.

This fi nding echoes a major point raised by Carneiro and Mylonakis (2) , namely, that without adjusting for spikes driven by disease publicity rather than the disease itself, Internet searches may lose much of their value in supplementing traditional surveillance measures. Our experience is that using Google Insights to survey a disease may not work well for diseases with considerable media exposure, in particular, emerging diseases such as pandemic (H1N1) 2009 or severe acute respiratory syndrome. Such outbreaks may require the more sophisticated approach used by Flu Trends, should it be extended to other diseases and more corners of the globe. We agree with Pelat et al. (3) that Google Insights may work well for less-publicized infectious diseases. The dividing line between well-publicized and unpublicized diseases may, however, remain ambiguous. Thus, to ensure that web search data refl ect disease incidence requires validation against traditional surveillance, although in that situation, the availability of corroborating traditional methods of surveillance limits the value of webquery data.

A.R.C. received research funding from the National University of Singapore. Other search patterns, such as for "swine fl u" and simplifi ed Chinese language terms for swine fl u and infl uenza, were similarly disassociated with actual disease incidence.

To the Editor: A scientifi c debate with public health implications wages: What caused the medieval European plague epidemics known as Black Death? Recent articles note inconsistencies between a rat fl eaborne pandemic of Yersinia pestis (the bacterium that causes bubonic plague) and the documented characteristics of Black Death (1, among others). Ayyadurai et al. (2) acknowledge that a rat fl ea-only hypothesis does not fi t Black Death observations, but they resolve theoretical transmission inconsistencies through a louse-borne hypothesis. Ayyadurai et al. base their surety of fact-that medieval "plagues" were caused by Y. pestis infection-on a 2007 study (3) in which 5 of 36 teeth of "plague" victims, none of which were dated to the Black Death era (1347-1351), contained biological evidence of Y. pestis. The 3 locations in that study were all port cities: 2 on the Mediterranean Sea and 1 on the Rhone River. As Duncan and Scott (4) note, bubonic plague most likely existed endemically near ship-borne trade, unlike the fast-moving epidemic fronts exhibited by medieval "plagues." Moreover, Gilbert et al. (5) found no Y. pestis DNA in 61 skeletons from primarily nonport locations in England, France, and Denmark.

We do not dispute the authors' claim that Y. pestis might have been present in some skeletons from port cities in France, or that body lice might, under certain circumstances, transmit the Orientalis biotype of Y. pestis; their work appears careful and considered. However, given the differences mentioned above and improved knowledge on the rapidity of virus mutation and worldwide transmission potential, we merely argue that the simplest explanation for medieval plagues has yet to be ruled out: that they may have resulted from a human-to-human transmitted virus. Adding complexity to an already complicated etiologic theory, and stating such as historical fact based on limited geography and sample size, does not seem congruent with Occam's razor. (1) reminded us of a littleknown paper (2) on rats and Black Death by our colleague and mentor David E. Davis. He researched and wrote in his retirement after years of research and refl ection on rat ecology and rodent-borne diseases (3, 4) . Rattus rattus is commonly recognized as the vertebrate host of fl ea-borne plague that swept through Europe in the 1300s, killing >50% of the population. Davis believed this explanation did not fi t what he knew of the ecologic requirements of fl eas and black rats. He studied reports of archeologic excavations and reviewed poems, medieval bestiaries, and paintings and concluded that these rats were scarce during the Black Death era.

His theory, based on historical information and investigative trips to Europe, was that invasive rats, if present, mostly occurred in low densities in port areas, not in rural inland areas. He noted that the expected rodent die-offs with bubonic plague were not associated with human epidemics and that rodent fl eas would not have been active during winter to transmit plague. Flea-borne transmission from rodents usually causes a few deaths per household, but deaths of entire households commonly occurred in the medieval epidemics. Human-tohuman transmission of pneumonic plague must have occurred, but as described by Ayyadurai et al., there was evidence of human bubonic plague, suggesting vector involvement. Davis did not present a viable reservoir/vector hypothesis for plague transmission; this and the later, well-known association of R. rattus and other rodents with plague throughout the world, may partially explain why his ideas received little attention. The fi nding that human body lice can be bubonic plague vectors suggests a mechanism for humanto-human transmission continuing during winter in inland areas and, as suggested by the authors, could also explain total deaths in households.

Epidemiology of infl uenza A(H1N1) virus infection in Japan

Offi ce of Pandemic Infl uenza Preparedness and Response, Ministry of Health, Labour and Welfare. Epidemiological information update: pandemic (H1N1)

Seasonal infl uenza in adults and childrendiagnosis, treatment, chemoprophylaxis, and institutional outbreak management: clinical practice guidelines of the Infectious Diseases Society of America

Infl uenza-associated hospitalizations in the United States

California Pandemic (H1N1) Working Group. Severe 2009 H1N1 infl uenza in pregnant and postpartum women in California

H1N1 infl uenza A and pregnancy outcomes in

Pandemic 2009 infl uenza A(H1N1) virus illness among pregnant women in the United States

H1N1 2009 infl uenza virus infection during pregnancy in the USA

Detecting infl uenza epidemics using search engine query data

Google Trends: a web-based tool for real-time surveillance of disease outbreaks

More diseases tracked by using Google Trends

Flu trends

Insights for search beta

Outbreak of pandemic infl uenza A (H1N1-2009) in Singapore

Real-time epidemic monitoring and forecasting of H1N1-2009 using infl uenza-like illness from general practice and family doctor clinics in Singapore

Validation of inverse seasonal peak mortality in medieval plagues, including the Black Death, in comparison to modern Yersinia pestis-variant diseases

Drancourt M. Body lice, Yersinia pestis Orientalis, and Black Death

Yersinia pestis Orientalis in remains of ancient plague patients

What caused the Black Death?

Absence of Yersinia pestis-specifi c DNA in human teeth from fi ve European excavations of putative plague victims