Neural network models for influenza forecasting with associated uncertainty using Web search activity trends.

Influenza affects millions of people every year. It causes a considerable amount of medical visits and hospitalisations as well as hundreds of thousands of deaths. Forecasting influenza prevalence with good accuracy can significantly help public health agencies to timely react to seasonal or novel strain epidemics. Although significant progress has been made, influenza forecasting remains a challenging modelling task. In this paper, we propose a methodological framework that improves over the state-of-the-art forecasting accuracy of influenza-like illness (ILI) rates in the United States. We achieve this by using Web search activity time series in conjunction with historical ILI rates as observations for training neural network (NN) architectures. The proposed models incorporate Bayesian layers to produce associated uncertainty intervals to their forecast estimates, positioning themselves as legitimate complementary solutions to more conventional approaches. The best performing NN, referred to as the iterative recurrent neural network (IRNN) architecture, reduces mean absolute error by 10.3% and improves skill by 17.1% on average in nowcasting and forecasting tasks across 4 consecutive flu seasons.

Using online search activity for earlier detection of gynaecological malignancy.

BACKGROUND: Ovarian cancer is the most lethal and endometrial cancer the most common gynaecological cancer in the UK, yet neither have a screening program in place to facilitate early disease detection. The aim is to evaluate whether online search data can be used to differentiate between individuals with malignant and benign gynaecological diagnoses. METHODS: This is a prospective cohort study evaluating online search data in symptomatic individuals (Google user) referred from primary care (GP) with a suspected cancer to a London Hospital (UK) between December 2020 and June 2022. Informed written consent was obtained and online search data was extracted via Google takeout and anonymised. A health filter was applied to extract health-related terms for 24 months prior to GP referral. A predictive model (outcome: malignancy) was developed using (1) search queries (terms model) and (2) categorised search queries (categories model). Area under the ROC curve (AUC) was used to evaluate model performance. 844 women were approached, 652 were eligible to participate and 392 were recruited. Of those recruited, 108 did not complete enrollment, 12 withdrew and 37 were excluded as they did not track Google searches or had an empty search history, leaving a cohort of 235. RESULTS: The cohort had a median age of 53 years old (range 20-81) and a malignancy rate of 26.0%. There was a difference in online search data between those with a benign and malignant diagnosis, noted as early as 360 days in advance of GP referral, when search queries were used directly, but only 60 days in advance, when queries were divided into health categories. A model using online search data from patients (n = 153) who performed health-related search and corrected for sample size, achieved its highest sample-corrected AUC of 0.82, 60 days prior to GP referral. CONCLUSIONS: Online search data appears to be different between individuals with malignant and benign gynaecological conditions, with a signal observed in advance of GP referral date. Online search data needs to be evaluated in a larger dataset to determine its value as an early disease detection tool and whether its use leads to improved clinical outcomes.

Google search patterns monitoring the daily health impact of heatwaves in England: How do the findings compare to established syndromic surveillance systems from 2013 to 2017?

One of the implications of climate change is a predicted increase in frequent and severe heatwaves. The impact of heatwaves on the health of the population is captured through real-time syndromic healthcare surveillance systems monitored daily in England during the summer months. Internet search data could potentially provide improved timeliness and help to assess the wider population health impact of heat by capturing a population sub-group who are symptomatic but do not seek healthcare. A retrospective observational study was carried out from June 2013 to September 2017 in England to compare daily trends in validated syndromic surveillance heat-related morbidity indicators against symptom-based heatwave related Google search terms. The degree of correlation was determined with Spearman correlation coefficients and lag assessment was carried out to determine timeliness. Daily increases in frequency in Google search terms during heatwave events correlated well with validated syndromic indicators. Correlation coefficients between search term frequency and syndromic indicators from 2013 to 2017 were highest with the telehealth service NHS 111 (range of 0.684-0.900 by search term). Lag analysis revealed a similar timeliness between the data sources, suggesting Google data did not provide a delayed or earlier signal in the context of England's syndromic surveillance systems. This work highlights the potential benefits for countries which lack established public health surveillance systems to monitor heat-related morbidity and the use of internet search data to assess the wider population health impact of exposure to heat.

Estimating the Population Impact of a New Pediatric Influenza Vaccination Program in England Using Social Media Content.

BACKGROUND: The rollout of a new childhood live attenuated influenza vaccine program was launched in England in 2013, which consisted of a national campaign for all 2 and 3 year olds and several pilot locations offering the vaccine to primary school-age children (4-11 years of age) during the influenza season. The 2014/2015 influenza season saw the national program extended to include additional pilot regions, some of which offered the vaccine to secondary school children (11-13 years of age) as well. OBJECTIVE: We utilized social media content to obtain a complementary assessment of the population impact of the programs that were launched in England during the 2013/2014 and 2014/2015 flu seasons. The overall community-wide impact on transmission in pilot areas was estimated for the different age groups that were targeted for vaccination. METHODS: A previously developed statistical framework was applied, which consisted of a nonlinear regression model that was trained to infer influenza-like illness (ILI) rates from Twitter posts originating in pilot (school-age vaccinated) and control (unvaccinated) areas. The control areas were then used to estimate ILI rates in pilot areas, had the intervention not taken place. These predictions were compared with their corresponding Twitter-based ILI estimates. RESULTS: Results suggest a reduction in ILI rates of 14% (1-25%) and 17% (2-30%) across all ages in only the primary school-age vaccine pilot areas during the 2013/2014 and 2014/2015 influenza seasons, respectively. No significant impact was observed in areas where two age cohorts of secondary school children were vaccinated. CONCLUSIONS: These findings corroborate independent assessments from traditional surveillance data, thereby supporting the ongoing rollout of the program to primary school-age children and providing evidence of the value of social media content as an additional syndromic surveillance tool.

Automatic identification of Web-based risk markers for health events.

BACKGROUND: The escalating cost of global health care is driving the development of new technologies to identify early indicators of an individual's risk of disease. Traditionally, epidemiologists have identified such risk factors using medical databases and lengthy clinical studies but these are often limited in size and cost and can fail to take full account of diseases where there are social stigmas or to identify transient acute risk factors. OBJECTIVE: Here we report that Web search engine queries coupled with information on Wikipedia access patterns can be used to infer health events associated with an individual user and automatically generate Web-based risk markers for some of the common medical conditions worldwide, from cardiovascular disease to sexually transmitted infections and mental health conditions, as well as pregnancy. METHODS: Using anonymized datasets, we present methods to first distinguish individuals likely to have experienced specific health events, and classify them into distinct categories. We then use the self-controlled case series method to find the incidence of health events in risk periods directly following a user's search for a query category, and compare to the incidence during other periods for the same individuals. RESULTS: Searches for pet stores were risk markers for allergy. We also identified some possible new risk markers; for example: searching for fast food and theme restaurants was associated with a transient increase in risk of myocardial infarction, suggesting this exposure goes beyond a long-term risk factor but may also act as an acute trigger of myocardial infarction. Dating and adult content websites were risk markers for sexually transmitted infections, such as human immunodeficiency virus (HIV). CONCLUSIONS: Web-based methods provide a powerful, low-cost approach to automatically identify risk factors, and support more timely and personalized public health efforts to bring human and economic benefits.

Detecting disease outbreaks in mass gatherings using Internet data.

BACKGROUND: Mass gatherings, such as music festivals and religious events, pose a health care challenge because of the risk of transmission of communicable diseases. This is exacerbated by the fact that participants disperse soon after the gathering, potentially spreading disease within their communities. The dispersion of participants also poses a challenge for traditional surveillance methods. The ubiquitous use of the Internet may enable the detection of disease outbreaks through analysis of data generated by users during events and shortly thereafter. OBJECTIVE: The intent of the study was to develop algorithms that can alert to possible outbreaks of communicable diseases from Internet data, specifically Twitter and search engine queries. METHODS: We extracted all Twitter postings and queries made to the Bing search engine by users who repeatedly mentioned one of nine major music festivals held in the United Kingdom and one religious event (the Hajj in Mecca) during 2012, for a period of 30 days and after each festival. We analyzed these data using three methods, two of which compared words associated with disease symptoms before and after the time of the festival, and one that compared the frequency of these words with those of other users in the United Kingdom in the days following the festivals. RESULTS: The data comprised, on average, 7.5 million tweets made by 12,163 users, and 32,143 queries made by 1756 users from each festival. Our methods indicated the statistically significant appearance of a disease symptom in two of the nine festivals. For example, cough was detected at higher than expected levels following the Wakestock festival. Statistically significant agreement (chi-square test, P<.01) between methods and across data sources was found where a statistically significant symptom was detected. Anecdotal evidence suggests that symptoms detected are indeed indicative of a disease that some users attributed to being at the festival. CONCLUSIONS: Our work shows the feasibility of creating a public health surveillance system for mass gatherings based on Internet data. The use of multiple data sources and analysis methods was found to be advantageous for rejecting false positives. Further studies are required in order to validate our findings with data from public health authorities.

Estimating the household secondary attack rate and serial interval of COVID-19 using social media.

We propose a method to estimate the household secondary attack rate (hSAR) of COVID-19 in the United Kingdom based on activity on the social media platform X, formerly known as Twitter. Conventional methods of hSAR estimation are resource intensive, requiring regular contact tracing of COVID-19 cases. Our proposed framework provides a complementary method that does not rely on conventional contact tracing or laboratory involvement, including the collection, processing, and analysis of biological samples. We use a text classifier to identify reports of people tweeting about themselves and/or members of their household having COVID-19 infections. A probabilistic analysis is then performed to estimate the hSAR based on the number of self or household, and self and household tweets of COVID-19 infection. The analysis includes adjustments for a reluctance of Twitter users to tweet about household members, and the possibility that the secondary infection was not acquired within the household. Experimental results for the UK, both monthly and weekly, are reported for the period from January 2020 to February 2022. Our results agree with previously reported hSAR estimates, varying with the primary variants of concern, e.g. delta and omicron. The serial interval (SI) is based on the time between the two tweets that indicate a primary and secondary infection. Experimental results, though larger than the consensus, are qualitatively similar. The estimation of hSAR and SI using social media data constitutes a new tool that may help in characterizing, forecasting and managing outbreaks and pandemics in a faster, affordable, and more efficient manner.

Providing early indication of regional anomalies in COVID-19 case counts in England using search engine queries.

Prior work has shown the utility of using Internet searches to track the incidence of different respiratory illnesses. Similarly, people who suffer from COVID-19 may query for their symptoms prior to accessing the medical system (or in lieu of it). To assist in the UK government's response to the COVID-19 pandemic we analyzed searches for relevant symptoms on the Bing web search engine from users in England to identify areas of the country where unexpected rises in relevant symptom searches occurred. These were reported weekly to the UK Health Security Agency to assist in their monitoring of the pandemic. Our analysis shows that searches for "fever" and "cough" were the most correlated with future case counts during the initial stages of the pandemic, with searches preceding case counts by up to 21 days. Unexpected rises in search patterns were predictive of anomalous rises in future case counts within a week, reaching an Area Under Curve of 0.82 during the initial phase of the pandemic, and later reducing due to changes in symptom presentation. Thus, analysis of regional searches for symptoms can provide an early indicator (of more than one week) of increases in COVID-19 case counts.

What on-line searches tell us about public interest and potential impact on behaviour in response to minimum unit pricing of alcohol in Scotland.

AIMS: To investigate whether the introduction of minimum unit pricing (MUP) in Scotland on 1 May 2018 was reflected in changes in the likelihood of alcohol-related queries submitted to an internet search engine, and in particular whether there was any evidence of increased interest in purchasing of alcohol from outside Scotland. DESIGN: Observational study in which individual queries to the internet Bing search engine for 2018 in Scotland and England were captured and analysed. Fluctuations over time in the likelihood of specific topic searches were examined. The patterns seen in Scotland were contrasted with those in England. SETTING: Scotland and England. PARTICIPANTS: People who used the Bing search engine during 2018. MEASUREMENTS: Numbers of daily queries submitted to Bing in 2018 on eight alcohol-related topics expressed as a proportion of queries on that day on any topic. These daily likelihoods were smoothed using a 14-day moving average for Scotland and England separately. FINDINGS: There were substantial peaks in queries about MUP itself, cheap sources of alcohol and online alcohol outlets at the time of introduction of MUP in May 2018 in Scotland, but not England. These were relatively short-lived. Queries related to intoxication and alcohol problems did not show a MUP peak, but were appreciably higher in Scotland than in England throughout 2018. CONCLUSIONS: Analysis of internet search engine queries appears to show that a fraction of people in Scotland may have considered circumventing minimum unit pricing in 2018 by looking for on-line alcohol retailers. The overall higher levels of queries related to alcohol problems in Scotland compared with England mirrors the corresponding differences in alcohol consumption and harms between the countries.

Tracking COVID-19 using online search.

Previous research has demonstrated that various properties of infectious diseases can be inferred from online search behaviour. In this work we use time series of online search query frequencies to gain insights about the prevalence of COVID-19 in multiple countries. We first develop unsupervised modelling techniques based on associated symptom categories identified by the United Kingdom's National Health Service and Public Health England. We then attempt to minimise an expected bias in these signals caused by public interest-as opposed to infections-using the proportion of news media coverage devoted to COVID-19 as a proxy indicator. Our analysis indicates that models based on online searches precede the reported confirmed cases and deaths by 16.7 (10.2-23.2) and 22.1 (17.4-26.9) days, respectively. We also investigate transfer learning techniques for mapping supervised models from countries where the spread of the disease has progressed extensively to countries that are in earlier phases of their respective epidemic curves. Furthermore, we compare time series of online search activity against confirmed COVID-19 cases or deaths jointly across multiple countries, uncovering interesting querying patterns, including the finding that rarer symptoms are better predictors than common ones. Finally, we show that web searches improve the short-term forecasting accuracy of autoregressive models for COVID-19 deaths. Our work provides evidence that online search data can be used to develop complementary public health surveillance methods to help inform the COVID-19 response in conjunction with more established approaches.

Risk factors, symptom reporting, healthcare-seeking behaviour and adherence to public health guidance: protocol for Virus Watch, a prospective community cohort study.

INTRODUCTION: The coronavirus (COVID-19) pandemic has caused significant global mortality and impacted lives around the world. Virus Watch aims to provide evidence on which public health approaches are most likely to be effective in reducing transmission and impact of the virus, and will investigate community incidence, symptom profiles and transmission of COVID-19 in relation to population movement and behaviours. METHODS AND ANALYSIS: Virus Watch is a household community cohort study of acute respiratory infections in England and Wales and will run from June 2020 to August 2021. The study aims to recruit 50 000 people, including 12 500 from minority ethnic backgrounds, for an online survey cohort and monthly antibody testing using home fingerprick test kits. Nested within this larger study will be a subcohort of 10 000 individuals, including 3000 people from minority ethnic backgrounds. This cohort of 10 000 people will have full blood serology taken between October 2020 and January 2021 and repeat serology between May 2021 and August 2021. Participants will also post self-administered nasal swabs for PCR assays of SARS-CoV-2 and will follow one of three different PCR testing schedules based on symptoms. ETHICS AND DISSEMINATION: This study has been approved by the Hampstead National Health Service (NHS) Health Research Authority Ethics Committee (ethics approval number 20/HRA/2320). We are monitoring participant queries and using these to refine methodology where necessary, and are providing summaries and policy briefings of our preliminary findings to inform public health action by working through our partnerships with our study advisory group, Public Health England, NHS and government scientific advisory panels.

Digital technologies in the public-health response to COVID-19.

Digital technologies are being harnessed to support the public-health response to COVID-19 worldwide, including population surveillance, case identification, contact tracing and evaluation of interventions on the basis of mobility data and communication with the public. These rapid responses leverage billions of mobile phones, large online datasets, connected devices, relatively low-cost computing resources and advances in machine learning and natural language processing. This Review aims to capture the breadth of digital innovations for the public-health response to COVID-19 worldwide and their limitations, and barriers to their implementation, including legal, ethical and privacy barriers, as well as organizational and workforce barriers. The future of public health is likely to become increasingly digital, and we review the need for the alignment of international strategies for the regulation, evaluation and use of digital technologies to strengthen pandemic management, and future preparedness for COVID-19 and other infectious diseases.

Relationship Between Media Coverage and Measles-Mumps-Rubella (MMR) Vaccination Uptake in Denmark: Retrospective Study.

BACKGROUND: Understanding the influence of media coverage upon vaccination activity is valuable when designing outreach campaigns to increase vaccination uptake. OBJECTIVE: To study the relationship between media coverage and vaccination activity of the measles-mumps-rubella (MMR) vaccine in Denmark. METHODS: We retrieved data on media coverage (1622 articles), vaccination activity (2 million individual registrations), and incidence of measles for the period 1997-2014. All 1622 news media articles were annotated as being provaccination, antivaccination, or neutral. Seasonal and serial dependencies were removed from the data, after which cross-correlations were analyzed to determine the relationship between the different signals. RESULTS: Most (65%) of the anti-vaccination media coverage was observed in the period 1997-2004, immediately before and following the 1998 publication of the falsely claimed link between autism and the MMR vaccine. There was a statistically significant positive correlation between the first MMR vaccine (targeting children aged 15 months) and provaccination media coverage (r=.49, P=.004) in the period 1998-2004. In this period the first MMR vaccine and neutral media coverage also correlated (r=.45, P=.003). However, looking at the whole period, 1997-2014, we found no significant correlations between vaccination activity and media coverage. CONCLUSIONS: Following the falsely claimed link between autism and the MMR vaccine, provaccination and neutral media coverage correlated with vaccination activity. This correlation was only observed during a period of controversy which indicates that the population is more susceptible to media influence when presented with diverging opinions. Additionally, our findings suggest that the influence of media is stronger on parents when they are deciding on the first vaccine of their children, than on the subsequent vaccine because correlations were only found for the first MMR vaccine.

The added value of online user-generated content in traditional methods for influenza surveillance.

There has been considerable work in evaluating the efficacy of using online data for health surveillance. Often comparisons with baseline data involve various squared error and correlation metrics. While useful, these overlook a variety of other factors important to public health bodies considering the adoption of such methods. In this paper, a proposed surveillance system that incorporates models based on recent research efforts is evaluated in terms of its added value for influenza surveillance at Public Health England. The system comprises of two supervised learning approaches trained on influenza-like illness (ILI) rates provided by the Royal College of General Practitioners (RCGP) and produces ILI estimates using Twitter posts or Google search queries. RCGP ILI rates for different age groups and laboratory confirmed cases by influenza type are used to evaluate the models with a particular focus on predicting the onset, overall intensity, peak activity and duration of the 2015/16 influenza season. We show that the Twitter-based models perform poorly and hypothesise that this is mostly due to the sparsity of the data available and a limited training period. Conversely, the Google-based model provides accurate estimates with timeliness of approximately one week and has the potential to complement current surveillance systems.

Optimizing hybrid spreading in metapopulations.

Epidemic spreading phenomena are ubiquitous in nature and society. Examples include the spreading of diseases, information, and computer viruses. Epidemics can spread by local spreading, where infected nodes can only infect a limited set of direct target nodes and global spreading, where an infected node can infect every other node. In reality, many epidemics spread using a hybrid mixture of both types of spreading. In this study we develop a theoretical framework for studying hybrid epidemics, and examine the optimum balance between spreading mechanisms in terms of achieving the maximum outbreak size. We show the existence of critically hybrid epidemics where neither spreading mechanism alone can cause a noticeable spread but a combination of the two spreading mechanisms would produce an enormous outbreak. Our results provide new strategies for maximising beneficial epidemics and estimating the worst outcome of damaging hybrid epidemics.

Estimating the secondary attack rate and serial interval of influenza-like illnesses using social media.

OBJECTIVES: Knowledge of the secondary attack rate (SAR) and serial interval (SI) of influenza is important for assessing the severity of seasonal epidemics of the virus. To date, such estimates have required extensive surveys of target populations. Here, we propose a method for estimating the intrafamily SAR and SI from postings on the Twitter social network. This estimate is derived from a large number of people reporting ILI symptoms in them and\or their immediate family members. DESIGN: We analyze data from the 2012-2013 and the 2013-2014 influenza seasons in England and find that increases in the estimated SAR precede increases in ILI rates reported by physicians. RESULTS: We hypothesize that observed variations in the peak value of SAR are related to the appearance of specific strains of the virus and demonstrate this by comparing the changes in SAR values over time in relation to known virology. In addition, we estimate SI (the average time between cases) as 2·41 days for 2012 and 2·48 days for 2013. CONCLUSIONS: The proposed method can assist health authorities by providing near-real-time estimation of SAR and SI, and especially in alerting to sudden increases thereof.

Applying informed coding and embedding to design a robust high-capacity watermark.

We describe a new watermarking system based on the principles of informed coding and informed embedding. This system is capable of embedding 1380 bits of information in images with dimensions 240 x 368 pixels. Experiments on 2000 images indicate the watermarks are robust to significant valumetric distortions, including additive noise, low-pass filtering, changes in contrast, and lossy compression. Our system encodes watermark messages with a modified trellis code in which a given message may be represented by a variety of different signals, with the embedded signal selected according to the cover image. The signal is embedded by an iterative method that seeks to ensure the message will not be confused with other messages, even after addition of noise. Fidelity is improved by the incorporation of perceptual shaping into the embedding process. We show that each of these three components improves performance substantially.