What Can Digital Disease Detection Learn from (an External Revision to) Google Flu Trends?
American Journal of Preventive MedicinePublished 2 July 2014
Mauricio Santillana, D. Wendong Zhang, Benjamin M. Althouse, John W. Ayers
Citations187
SJR quartileQ1
SJR score2.08
SNIP1.50
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
GFT may be inaccurate, but improved methodologic underpinnings can yield accurate predictions, and applying similar methods elsewhere can improve digital disease detection, with broader transparency, improved accuracy, and real-world public health impacts.
Abstract
GFT may be inaccurate, but improved methodologic underpinnings can yield accurate predictions. Applying similar methods elsewhere can improve digital disease detection, with broader transparency, improved accuracy, and real-world public health impacts.
Keywords
MathematicsMedicine
Journal of the Royal Statistical Society Series B (Statistical Methodology)Regression Shrinkage and Selection Via the Lasso
51,790 Citations1996Robert Tibshirani
A new method for estimation in linear models called the lasso, which minimizes the residual sum of squares subject to the sum of the absolute value of the coefficients being less than a constant, is proposed.
The Elements of Statistical Learning: Data Mining, Inference, and Prediction
19,345 Citations2013Trevor Hastie, Robert Tibshirani +1 more
NatureDetecting influenza epidemics using search engine query data
4,406 Citations2008Jeremy Ginsberg, Matthew H. Mohebbi +4 more
A method of analysing large numbers of Google search queries to track influenza-like illness in a population and accurately estimate the current level of weekly influenza activity in each region of the United States with a reporting lag of about one day is presented.
ScienceThe Parable of Google Flu: Traps in Big Data Analysis
2,520 Citations2014David Lazer, Ryan Kennedy +2 more
Large errors in flu prediction were largely avoidable, which offers lessons for the use of big data.
RPK (Politechniki Krakowskiej)Artificial Intelligence: A Modern Approach
1,746 Citations1995Stuart Russell, Peter Norvig
Proceedings of the International AAAI Conference on Web and Social MediaYou Are What You Tweet: Analyzing Twitter for Public Health
964 Citations2021Michael Paul, Mark Dredze
This work applies the recently introduced Ailment Topic Aspect Model to over one and a half million health related tweets and discovers mentions of over a dozen ailments, including allergies, obesity and insomnia, suggesting that Twitter has broad applicability for public health research.
Advances in geophysicsData Assimilation in Meteorology and Oceanography
933 Citations1991Michael Ghil, Paola Malanotte‐Rizzoli
New England Journal of MedicineDigital Disease Detection — Harnessing the Web for Public Health Surveillance
837 Citations2009John S. Brownstein, Clark C. Freifeld +1 more
A new generation of disease-surveillance “mashups” can mine, categorize, filter, and visualize online intelligence about epidemics in real time.
American Journal of Tropical Medicine and HygieneSocial and News Media Enable Estimation of Epidemiological Patterns Early in the 2010 Haitian Cholera Outbreak
445 Citations2012Rumi Chunara, Jason R. Andrews +1 more
This work assessed correlation of volume of cholera-related HealthMap news media reports, Twitter postings, and governmentCholera cases reported in the first 100 days of the 2010 Haitian cholERA outbreak to find trends in volume of informal sources significantly correlated in time with official case data and was available up to 2 weeks earlier.
Proceedings of the National Academy of SciencesForecasting seasonal outbreaks of influenza
440 Citations2012Jeffrey Shaman, Alicia Karspeck
It is indicated that real-time skillful predictions of peak timing can be made more than 7 wk in advance of the actual peak, and confidence in those predictions can be inferred from the spread of the forecast ensemble.
PubMedInfodemiology: tracking flu-related searches on the web for syndromic surveillance.
420 Citations2006Günther Eysenbach
Systematically collecting and analyzing health information demand data from the Internet has considerable potential to be used for syndromic surveillance.
American Journal of Preventive MedicineInfodemiology and Infoveillance
417 Citations2011Günther Eysenbach
PLoS ONEAssessing Google Flu Trends Performance in the United States during the 2009 Influenza Virus A (H1N1) Pandemic
374 Citations2011S. M. Cook, Corrie Conrad +2 more
Both GFT models performed well prior to and during pH1N1, although the updated model performed better during pH0.1, especially during the summer months, and changes in health-seeking behavior may have played a part.
PLoS Computational BiologyReassessing Google Flu Trends Data for Detection of Seasonal and Pandemic Influenza: A Comparative Epidemiological Study at Three Geographic Scales
323 Citations2013Donald R. Olson, Kevin Konty +3 more
GFT data may not provide reliable surveillance for seasonal or pandemic influenza and should be interpreted with caution until the algorithm can be improved and evaluated, and current internet search query data are no substitute for timely local clinical and laboratory surveillance, or national surveillance based on local data collection.
American Journal of Preventive MedicineTracking the Rise in Popularity of Electronic Nicotine Delivery Systems (Electronic Cigarettes) Using Search Query Surveillance
301 Citations2011John W. Ayers, Kurt M. Ribisl +1 more
The online popularity of ENDS has surpassed that of snus and NRTs, which have been on the market for far longer, and is quickly outpacing Chantix or Champix, and suggests ENDS are used to bypass, or quit in response to, smoking restrictions.
PLoS neglected tropical diseasesUsing Web Search Query Data to Monitor Dengue Epidemics: A New Model for Neglected Tropical Disease Surveillance
279 Citations2011Emily H. Chan, Vikram Sahai +2 more
Web search query data were found to be capable of tracking dengue activity in Bolivia, Brazil, India, Indonesia and Singapore, and represent valuable complement to assist with traditional d Dengue surveillance.
PLoS neglected tropical diseasesPrediction of Dengue Incidence Using Search Query Surveillance
274 Citations2011Benjamin M. Althouse, Yih Yng Ng +1 more
Internet search terms predict incidence and periods of large incidence of dengue with high accuracy and may prove useful in areas with underdeveloped surveillance systems.
PLoS ONEMonitoring Influenza Epidemics in China with Search Query from Baidu
252 Citations2013Qingyu Yuan, Elaine O. Nsoesie +4 more
This paper presents a method that uses Internet search query data from Baidu to model and monitor influenza activity in China and presents a comprehensive technique for keyword selection, keyword filtering, index composition and modeling and detection of influenzaActivity in China.
Clinical Infectious DiseasesGoogle Flu Trends: Correlation With Emergency Department Influenza Rates and Crowding Metrics
234 Citations2012Andrea Dugas, Yu‐Hsiang Hsieh +7 more
City-level GFT shows strong correlation with influenza cases and ED ILI visits, validating its use as an ED surveillance tool.
Canadian Medical Association JournalEarly detection of disease outbreaks using the Internet
231 Citations2009Kumanan Wilson, John S. Brownstein
A large number of countries lack surveillance capacity for rapid and effective detection of infectious disease outbreaks, and many countries lack the public health infrastructure to support such detection.
Canadian Medical Association JournalEarly detection of disease outbreaks using the Internet
217 Citations2009Kumanan Wilson, John S. Brownstein
American Journal of Preventive MedicineSeasonality in Seeking Mental Health Information on Google
207 Citations2013John W. Ayers, Benjamin M. Althouse +3 more
These are the first data published on patterns of seasonality in information seeking encompassing all the major mental illnesses, notable also because they likely would have gone undetected using traditional surveillance.
EurosurveillanceInterim analysis of pandemic influenza (H1N1) 2009 in Australia: surveillance trends, age of infection and effectiveness of seasonal vaccination
193 Citations2009H Kelly, Kristina A Grant
There was no evidence of significant protection from seasonal vaccine against pandemic influenza virus infection in any age group and at the time of reporting, the peak ILI activity in 2009 had been reached.
PLoS ONEUsing Google Trends for Influenza Surveillance in South China
179 Citations2013Min Kang, Haojie Zhong +3 more
This study shows that Google Trends in Chinese can be used as a complementary source of data for influenza surveillance in south China and develops new models using search trends in Chinese language to estimate local disease activity and detect early signals of outbreaks.
PLoS ONEMonitoring Influenza Activity in the United States: A Comparison of Traditional Surveillance Systems with Google Flu Trends
149 Citations2011Justin R. Ortiz, Hong Zhou +4 more
It is demonstrated that while Google Flu Trends is highly correlated with rates of ILI, it has a lower correlation with surveillance for laboratory-confirmed influenza.
Journal of Affective DisordersAssociation of Internet search trends with suicide death in Taipei City, Taiwan, 2004–2009
134 Citations2011Albert C. Yang, Shi-Jen Tsai +2 more
Results show that a set of suicide-related search terms, the trends of which either temporally coincided or preceded trends of suicide data, were associated with suicide death.
JAMACould Behavioral Medicine Lead the Web Data Revolution?
131 Citations2014John W. Ayers, Benjamin M. Althouse +1 more
Web data are potentially the only source for real-time insights into behavioral medicine, where web data can be available almost immediately compared to a 365-day lag time between annual surveys, and can be an important source for identifying new hypotheses.
Proceedings of the National Academy of SciencesData assimilation and its applications
104 Citations2000Bin Wang, Xiaolei Zou +1 more
Three examples are presented to illustrate the use of new types of observations and the ability of improving forecasting or modeling in studies of atmospheric and oceanic problems.
UrologyInternet Search Trends Analysis Tools Can Provide Real-time Data on Kidney Stone Disease in the United States
75 Citations2011Scott Willard, Mike M. Nguyen
The search term ranking suggested that Internet users are most interested in the diagnosis, followed by etiology, infections, and treatment, which appears to be accurately reflected in Internet search trends data.
UrologyUse of Google Insights for Search to Track Seasonal and Geographic Kidney Stone Incidence in the United States
71 Citations2011Benjamin N. Breyer, Śaunak Sen +4 more
Internet search volume activity for kidney stones correlates with temporal and regional kidney stone insurance claims data and has the potential to serve as a surrogate for kidney stone incidence.
Emerging infectious diseasesDiseases Tracked by Using Google Trends, Spain
71 Citations2009Antonio Valdivia, Susana Monge
The utility of Internet queries for the surveillance of ILI and chickenpox in Spain can be used as an early warning tool used complementarily to standard surveillance systems.
Malaria JournalUsing search queries for malaria surveillance, Thailand
70 Citations2013Alex Ocampo, Rumi Chunara +1 more
Journal of Affective DisordersNovel surveillance of psychological distress during the great recession
68 Citations2012John W. Ayers, Benjamin M. Althouse +6 more
A snapshot of recent associations between the economy and PD queries is provided but also a framework and toolkit for real-time surveillance going forward so health resources, clinician screening patterns, and policy debate may be informed by changes in PD query trends.
Preventive MedicineDo celebrity cancer diagnoses promote primary cancer prevention?
57 Citations2013John W. Ayers, Benjamin M. Althouse +2 more
Discovery of this dynamic suggests the public should be further encouraged to consider primary cancer prevention around celebrity diagnoses, though more cases, cancers, and prevention behaviors must be explored.
American Journal of Preventive MedicinePopulation Health Concerns During the United States’ Great Recession
56 Citations2014Benjamin M. Althouse, Jon-Patrick Allem +3 more
Google queries indicate that the Great Recession coincided with substantial increases in health concerns, hinting at how population health specifically changed during that time.
Tobacco ControlMonitoring of non-cigarette tobacco use using Google Trends
55 Citations2014Patricia Cavazos‐Rehg, Melissa J. Krauss +5 more
This study validate that Google Trends has the potential to be a valuable monitoring tool for tobacco use and near real-time monitoring features of Google Trends may complement traditional surveillance methods and lead to faster and more convenient monitoring of emerging trends in tobacco use.
PLoS ONEComparison: Flu Prescription Sales Data from a Retail Pharmacy in the US with Google Flu Trends and US ILINet (CDC) Data as Flu Activity Indicator
54 Citations2012Avinash Patwardhan, Robert Bilkovski
This study compared prescription sales data from a large drug retailing pharmacy chain in the United States with Google Flu trends surveillance system data as a flu activity indicator and found that the two were highly correlated.
JAMA Internal MedicineCircaseptan (Weekly) Rhythms in Smoking Cessation Considerations
54 Citations2013John W. Ayers, Benjamin M. Althouse +2 more
Smokers’ rhythms in contemplating quitting or making quit attempts are poorly understood, but weekly “circaseptan” time-cycles may likewise exist.
GOOGLE DISEASE TRENDS: AN UPDATE
47 Citations2013Patrick Copeland, Raquel Romano +4 more
In addition to the general public, an important target audience for GFT has been public health officials, who can benefit from reliable daily estimates and often make far-reaching decisions based on predicted flu incidence.
Nicotine & Tobacco ResearchDigital Detection for Tobacco Control: Online Reactions to the 2009 U.S. Cigarette Excise Tax Increase
35 Citations2013John W. Ayers, Benjamin M. Althouse +2 more
This strategy can support evaluations that temporally link tobacco control measures with instantaneous population reactions, as well as serve as a springboard for traditional studies, for example, including survey questionnaire design.
Journal of Medical Internet ResearchGuess Who’s Not Coming to Dinner? Evaluating Online Restaurant Reservations for Disease Surveillance
34 Citations2014Elaine O. Nsoesie, David L. Buckeridge +1 more
This study introduces and demonstrates the potential value of restaurant use data for event surveillance and noted similarities in time series trends between data on influenza activity and restaurant use.
Emerging infectious diseasesInternet Queries and Methicillin-Resistant<i>Staphylococcus aureus</i>Surveillance
27 Citations2011Vanja Dukić, Michael David +1 more
Lecture notes in electrical engineeringRecent Progress in Data Engineering and Internet Technology
16 Citations2012Ford Lumban Gaol
The design of Robust Fault Detection Filter of T-S Fuzzy Time-delay Systems and an eigenvector-based kernel clustering approach to detecting communities in complex networks are presented.
Lecture notes in electrical engineeringRecent Progress in Data Engineering and Internet Technology
13 Citations2012Ford Lumban Gaol
The Open Public Health JournalCorrelation of ”Google Flu Trends“ with Sentinel Surveillance Data forInfluenza in 2009 in Japan
8 Citations2011Koji Wada, Hiroshi Ohta +1 more
Google Flu Trends (GFT) aggregates Google search data to estimate flu activity in 28 countries and shows people's interest in influenza affects GFT data, which could reflect influenza spread in mega-cities such as Tokyo.
Lecture notes in electrical engineeringInfluenza Epidemics Detection Based on Google Search Queries
4 Citations2012Fan Liu, Benfu Lv +2 more
A hybrid model with dynamic search query set is developed, which was more accurate in influenza forecast than Google flu trends, especially for the irregular new influenza strain forecasts.
Development of a Real-Time Estimate of Flu Activity in the United States Using Dynamically Updated Lasso Regressions and Google Search Queries
2 Citations2013Wendong Zhang
Results show that improving the underlying regression model makes substantial and long-term improvements to ILI surveillance using query-based methods, rendering the update to GFT’s database of search queries unnecessary.