Combining Participatory Influenza Surveillance with Modeling and Forecasting: Three Alternative Approaches

Integrating information from historical data into mechanistic models for influenza forecasting

Seasonal influenza causes significant annual morbidity and mortality worldwide. In France, it is estimated that, on average, 2 million individuals consult their GP for influenza-like-illness (ILI) every year. Traditionally, mathematical models used for epidemic forecasting can either include parameters capturing the infection process (mechanistic or compartmental models) or rely on time series analysis approaches that do not make mechanistic assumptions (statistical or phenomenological models). While the latter make extensive use of past epidemic data, mechanistic models are usually independently initialized in each season. As a result, forecasts from such models can contain trajectories that are vastly different from past epidemics. We developed a mechanistic model that takes into account epidemic data from training seasons when producing forecasts. The parameters of the model are estimated via a first particle filter running on the observed data. A second particle filter is then used to produce forecasts compatible with epidemic trajectories from the training set. The model was calibrated and tested on 35 years' worth of surveillance data from the French Sentinelles Network, representing the weekly number of patients consulting for ILI over the period 1985-2019. Our results show that the new method improves upon standard mechanistic approaches. In particular, when retrospectively tested on the available data, our model provides increased accuracy for short-term forecasts (from one to four weeks into the future) and peak timing and intensity. Our new approach for epidemic forecasting allows the integration of key strengths of the statistical approach into the mechanistic modelling framework and represents an attempt to provide accurate forecasts by making full use of the rich surveillance dataset collected in France since 1985.

Comparison of Synthetic Data Generation Techniques for Control Group Survival Data in Oncology Clinical Trials: Simulation Study

Background

Synthetic patient data (SPD) generation for survival analysis in oncology trials holds significant potential for accelerating clinical development. Various machine learning methods, including classification and regression trees (CART), random forest (RF), Bayesian network (BN), and conditional tabular generative adversarial network (CTGAN), have been used for this purpose, but their performance in reflecting actual patient survival data remains under investigation.

Objective

The aim of this study was to determine the most suitable SPD generation method for oncology trials, specifically focusing on both progression-free survival (PFS) and overall survival (OS), which are the primary evaluation end points in oncology trials. To achieve this goal, we conducted a comparative simulation of 4 generation methods, including CART, RF, BN, and the CTGAN, and the performance of each method was evaluated.

Methods

Using multiple clinical trial data sets, 1000 data sets were generated by using each method for each clinical trial data set and evaluated as follows: (1) median survival time (MST) of PFS and OS; (2) hazard ratio distance (HRD), which indicates the similarity between the actual survival function and a synthetic survival function; and (3) visual analysis of Kaplan-Meier (KM) plots. Each method's ability to mimic the statistical properties of real patient data was evaluated from these multiple angles.

Results

In most simulation cases, CART demonstrated the high percentages of MSTs for synthetic data falling within the 95% CI range of the MST of the actual data. These percentages ranged from 88.8% to 98.0% for PFS and from 60.8% to 96.1% for OS. In the evaluation of HRD, CART revealed that HRD values were concentrated at approximately 0.9. Conversely, for the other methods, no consistent trend was observed for either PFS or OS. CART demonstrated better similarity than RF, in that CART caused overfitting and RF (a kind of ensemble learning approach) prevented it. In SPD generation, the statistical properties close to the actual data should be the focus, not a well-generalized prediction model. Both the BN and CTGAN methods cannot accurately reflect the statistical properties of the actual data because small data sets are not suitable.

Conclusions

As a method for generating SPD for survival data from small data sets, such as clinical trial data, CART demonstrated to be the most effective method compared to RF, BN, and CTGAN. Additionally, it is possible to improve CART-based generation methods by incorporating feature engineering and other methods in future work.

Long-Term Regional Influenza-Like-Illness Forecasting Using Exogenous Data

Disease forecasting is a longstanding problem for the research community, which aims at informing and improving decisions with the best available evidence. Specifically, the interest in respiratory disease forecasting has dramatically increased since the beginning of the coronavirus pandemic, rendering the accurate prediction of influenza-like-illness (ILI) a critical task. Although methods for short-term ILI forecasting and nowcasting have achieved good accuracy, their performance worsens at long-term ILI forecasts. Machine learning models have outperformed conventional forecasting approaches enabling to utilize diverse exogenous data sources, such as social media, internet users' search query logs, and climate data. However, the most recent deep learning ILI forecasting models use only historical occurrence data achieving state-of-the-art results. Inspired by recent deep neural network architectures in time series forecasting, this work proposes the Regional Influenza-Like-Illness Forecasting (ReILIF) method for regional long-term ILI prediction. The proposed architecture takes advantage of diverse exogenous data, that are, meteorological and population data, introducing an efficient intermediate fusion mechanism to combine the different types of information with the aim to capture the variations of ILI from various views. The efficacy of the proposed approach compared to state-of-the-art ILI forecasting methods is confirmed by an extensive experimental study following standard evaluation measures.

Real-time estimation of the effective reproduction number of COVID-19 from behavioral data

Monitoring the effective reproduction number [Formula: see text] of a rapidly unfolding pandemic in real-time is key to successful mitigation and prevention strategies. However, existing methods based on case numbers, hospital admissions or fatalities suffer from multiple measurement biases and temporal lags due to high test positivity rates or delays in symptom development or administrative reporting. Alternative methods such as web search and social media tracking are less directly indicating epidemic prevalence over time. We instead record age-stratified anonymous contact matrices at a daily resolution using a longitudinal online-offline survey in Hungary during the first two waves of the COVID-19 pandemic. This approach is innovative, cheap, and provides information in near real-time for estimating [Formula: see text] at a daily resolution. Moreover, it allows to complement traditional surveillance systems by signaling periods when official monitoring infrastructures are unreliable due to observational biases.

Enabling Multicentric Participatory Disease Surveillance for Global Health Enhancement: Viewpoint on Global Flu View

Participatory surveillance (PS) has been defined as the bidirectional process of transmitting and receiving data for action by directly engaging the target population. Often represented as self-reported symptoms directly from the public, PS can provide evidence of an emerging disease or concentration of symptoms in certain areas, potentially identifying signs of an early outbreak. The construction of sets of symptoms to represent various disease syndromes provides a mechanism for the early detection of multiple health threats. Global Flu View (GFV) is the first-ever system that merges influenza-like illness (ILI) data from more than 8 countries plus 1 region (Hong Kong) on 4 continents for global monitoring of this annual health threat. GFV provides a digital ecosystem for spatial and temporal visualization of syndromic aggregates compatible with ILI from the various systems currently participating in GFV in near real time, updated weekly. In 2018, the first prototype of a digital platform to combine data from several ILI PS programs was created. At that time, the priority was to have a digital environment that brought together different programs through an application program interface, providing a real time map of syndromic trends that could demonstrate where and when ILI was spreading in various regions of the globe. After 2 years running as an experimental model and incorporating feedback from partner programs, GFV was restructured to empower the community of public health practitioners, data scientists, and researchers by providing an open data channel among these contributors for sharing experiences across the network. GFV was redesigned to serve not only as a data hub but also as a dynamic knowledge network around participatory ILI surveillance by providing knowledge exchange among programs. Connectivity between existing PS systems enables a network of cooperation and collaboration with great potential for continuous public health impact. The exchange of knowledge within this network is not limited only to health professionals and researchers but also provides an opportunity for the general public to have an active voice in the collective construction of health settings. The focus on preparing the next generation of epidemiologists will be of great importance to scale innovative approaches like PS. GFV provides a useful example of the value of globally integrated PS data to help reduce the risks and damages of the next pandemic.

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.

Crime, inequality and public health: a survey of emerging trends in urban data science

Urban agglomerations are constantly and rapidly evolving ecosystems, with globalization and increasing urbanization posing new challenges in sustainable urban development well summarized in the United Nations' Sustainable Development Goals (SDGs). The advent of the digital age generated by modern alternative data sources provides new tools to tackle these challenges with spatio-temporal scales that were previously unavailable with census statistics. In this review, we present how new digital data sources are employed to provide data-driven insights to study and track (i) urban crime and public safety; (ii) socioeconomic inequalities and segregation; and (iii) public health, with a particular focus on the city scale.

Calculating incidence of Influenza-like and COVID-like symptoms from Flutracking participatory survey data

This article describes a new method for estimating weekly incidence (new onset) of symptoms consistent with Influenza and COVID-19, using data from the Flutracking survey. The method mitigates some of the known self-selection and symptom-reporting biases present in existing approaches to this type of participatory longitudinal survey data. The key novel steps in the analysis are: 1) Identifying new onset of symptoms for three different Symptom Groupings: COVID-like illness (CLI1+, CLI2+), and Influenza-like illness (ILI), for responses reported in the Flutracking survey. 2) Adjusting for symptom reporting bias by restricting the analysis to a sub-set of responses from those participants who have consistently responded for a number of weeks prior to the analysis week. 3) Weighting responses by age to adjust for self-selection bias in order to account for the under- and over-representation of different age groups amongst the survey participants. This uses the survey package [22] in R [30]. 4) Constructing 95% point-wise confidence bands for incidence estimates using weighted logistic regression from the survey package [21] in R [28]. In addition to describing these steps, the article demonstrates an application of this method to Flutracking data for the 12 months from 27th April 2020 until 25th April 2021.

Innovative Use of a Mobile Web Application to Remotely Monitor Nonhospitalized Patients with COVID-19

Introduction: Most patients with COVID-19 do not require hospitalization but may need close monitoring, which can strain primary care practices. Our objective was to describe the implementation of a mobile web application to monitor COVID-19 signs and symptoms among nonhospitalized primary care patients and to assess the feasibility and acceptability of the application. Study Design: Retrospective analysis of (1) mobile web application data from March through December 2020 and (2) cross-sectional surveys administered in June 2020. Materials and Methods: We enrolled nonhospitalized patients and staff from nine New England primary care practices across 29 sites. Outcomes included feasibility and acceptability of the application as measured by the proportion of texts that resulted in a response, proportion of patients who agreed using the application was easy, and proportion of practice staff who agreed the application reduced outreach burden and that they would recommend use. Results: Five thousand five hundred thirty-two patients used the mobile web application, with 26,466 total responses. Overall, 78% of the daily texts resulted in a response from patients. Most patients agreed that responding to texts was easy (95%) and that they would be willing to participate in other texting programs (78%). Most staff agreed that the program reduced burden of outreach (94%) and that they would recommend use to other practices (100%). Conclusions: Use of a COVID-19 symptom tracking application was feasible and acceptable to patients and primary care practice staff. Outpatient practices should consider use of mobile web applications to monitor nonhospitalized patients with other acute illnesses.

Comparison of longitudinal trends in self-reported symptoms and COVID-19 case activity in Ontario, Canada

BACKGROUND

Limitations in laboratory diagnostic capacity impact population surveillance of COVID-19. It is currently unknown whether participatory surveillance tools for COVID-19 correspond to government-reported case trends longitudinally and if it can be used as an adjunct to laboratory testing. The primary objective of this study was to determine whether self-reported COVID-19-like illness reflected laboratory-confirmed COVID-19 case trends in Ontario Canada.

METHODS

We retrospectively analyzed longitudinal self-reported symptoms data collected using an online tool-Outbreaks Near Me (ONM)-from April 20th, 2020, to March 7th, 2021 in Ontario, Canada. We measured the correlation between COVID-like illness among respondents and the weekly number of PCR-confirmed COVID-19 cases and provincial test positivity. We explored contemporaneous changes in other respiratory viruses, as well as the demographic characteristics of respondents to provide context for our findings.

RESULTS

Between 3,849-11,185 individuals responded to the symptom survey each week. No correlations were seen been self-reported CLI and either cases or test positivity. Strong positive correlations were seen between CLI and both cases and test positivity before a previously documented rise in rhinovirus/enterovirus in fall 2020. Compared to participatory surveillance respondents, a higher proportion of COVID-19 cases in Ontario consistently came from low-income, racialized and immigrant areas of the province- these groups were less well represented among survey respondents.

INTERPRETATION

Although digital surveillance systems are low-cost tools that have been useful to signal the onset of viral outbreaks, in this longitudinal comparison of self-reported COVID-like illness to Ontario COVID-19 case data we did not find this to be the case. Seasonal respiratory virus transmission and population coverage may explain this discrepancy.

Digital SARS-CoV-2 Detection Among Hospital Employees: Participatory Surveillance Study

Background

The implementation of novel techniques as a complement to traditional disease surveillance systems represents an additional opportunity for rapid analysis.

Objective

The objective of this work is to describe a web-based participatory surveillance strategy among health care workers (HCWs) in two Swiss hospitals during the first wave of COVID-19.

Methods

A prospective cohort of HCWs was recruited in March 2020 at the Cantonal Hospital of St. Gallen and the Eastern Switzerland Children’s Hospital. For data analysis, we used a combination of the following techniques: locally estimated scatterplot smoothing (LOESS) regression, Spearman correlation, anomaly detection, and random forest.

Results

From March 23 to August 23, 2020, a total of 127,684 SMS text messages were sent, generating 90,414 valid reports among 1004 participants, achieving a weekly average of 4.5 (SD 1.9) reports per user. The symptom showing the strongest correlation with a positive polymerase chain reaction test result was loss of taste. Symptoms like red eyes or a runny nose were negatively associated with a positive test. The area under the receiver operating characteristic curve showed favorable performance of the classification tree, with an accuracy of 88% for the training data and 89% for the test data. Nevertheless, while the prediction matrix showed good specificity (80.0%), sensitivity was low (10.6%).

Conclusions

Loss of taste was the symptom that was most aligned with COVID-19 activity at the population level. At the individual level—using machine learning–based random forest classification—reporting loss of taste and limb/muscle pain as well as the absence of runny nose and red eyes were the best predictors of COVID-19.

Cross-sectional telephone surveys as a tool to study epidemiological factors and monitor seasonal influenza activity in Malta

Background

Seasonal influenza has major implications for healthcare services as outbreaks often lead to high activity levels in health systems. Being able to predict when such outbreaks occur is vital. Mathematical models have extensively been used to predict epidemics of infectious diseases such as seasonal influenza and to assess effectiveness of control strategies. Availability of comprehensive and reliable datasets used to parametrize these models is limited. In this paper we combine a unique epidemiological dataset collected in Malta through General Practitioners (GPs) with a novel method using cross-sectional surveys to study seasonal influenza dynamics in Malta in 2014–2016, to include social dynamics and self-perception related to seasonal influenza.

Methods

Two cross-sectional public surveys (n = 406 per survey) were performed by telephone across the Maltese population in 2014–15 and 2015–16 influenza seasons. Survey results were compared with incidence data (diagnosed seasonal influenza cases) collected by GPs in the same period and with Google Trends data for Malta. Information was collected on whether participants recalled their health status in past months, occurrences of influenza symptoms, hospitalisation rates due to seasonal influenza, seeking GP advice, and other medical information.

Results

We demonstrate that cross-sectional surveys are a reliable alternative data source to medical records. The two surveys gave comparable results, indicating that the level of recollection among the public is high. Based on two seasons of data, the reporting rate in Malta varies between 14 and 22%. The comparison with Google Trends suggests that the online searches peak at about the same time as the maximum extent of the epidemic, but the public interest declines and returns to background level. We also found that the public intensively searched the Internet for influenza-related terms even when number of cases was low.

Conclusions

Our research shows that a telephone survey is a viable way to gain deeper insight into a population’s self-perception of influenza and its symptoms and to provide another benchmark for medical statistics provided by GPs and Google Trends. The information collected can be used to improve epidemiological modelling of seasonal influenza and other infectious diseases, thus effectively contributing to public health.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-11862-x.

Digital Surveillance Through an Online Decision Support Tool for COVID-19 Over One Year of the Pandemic in Italy: Observational Study

BACKGROUND

Italy has experienced severe consequences (ie, hospitalizations and deaths) during the COVID-19 pandemic. Online decision support systems (DSS) and self-triage applications have been used in several settings to supplement health authority recommendations to prevent and manage COVID-19. A digital Italian health tech startup, Paginemediche, developed a noncommercial, online DSS with a chat user interface to assist individuals in Italy manage their potential exposure to COVID-19 and interpret their symptoms since early in the pandemic.

OBJECTIVE

This study aimed to compare the trend in online DSS sessions with that of COVID-19 cases reported by the national health surveillance system in Italy, from February 2020 to March 2021.

METHODS

We compared the number of sessions by users with a COVID-19-positive contact and users with COVID-19-compatible symptoms with the number of cases reported by the national surveillance system. To calculate the distance between the time series, we used the dynamic time warping algorithm. We applied Symbolic Aggregate approXimation (SAX) encoding to the time series in 1-week periods. We calculated the Hamming distance between the SAX strings. We shifted time series of online DSS sessions 1 week ahead. We measured the improvement in Hamming distance to verify the hypothesis that online DSS sessions anticipate the trends in cases reported to the official surveillance system.

RESULTS

We analyzed 75,557 sessions in the online DSS; 65,207 were sessions by symptomatic users, while 19,062 were by contacts of individuals with COVID-19. The highest number of online DSS sessions was recorded early in the pandemic. Second and third peaks were observed in October 2020 and March 2021, respectively, preceding the surge in notified COVID-19 cases by approximately 1 week. The distance between sessions by users with COVID-19 contacts and reported cases calculated by dynamic time warping was 61.23; the distance between sessions by symptomatic users was 93.72. The time series of users with a COVID-19 contact was more consistent with the trend in confirmed cases. With the 1-week shift, the Hamming distance between the time series of sessions by users with a COVID-19 contact and reported cases improved from 0.49 to 0.46. We repeated the analysis, restricting the time window to between July 2020 and December 2020. The corresponding Hamming distance was 0.16 before and improved to 0.08 after the time shift.

CONCLUSIONS

Temporal trends in the number of online COVID-19 DSS sessions may precede the trend in reported COVID-19 cases through traditional surveillance. The trends in sessions by users with a contact with COVID-19 may better predict reported cases of COVID-19 than sessions by symptomatic users. Data from online DSS may represent a useful supplement to traditional surveillance and support the identification of early warning signals in the COVID-19 pandemic.

Estimating the cumulative incidence of COVID-19 in the United States using influenza surveillance, virologic testing, and mortality data: Four complementary approaches

Effectively designing and evaluating public health responses to the ongoing COVID-19 pandemic requires accurate estimation of the prevalence of COVID-19 across the United States (US). Equipment shortages and varying testing capabilities have however hindered the usefulness of the official reported positive COVID-19 case counts. We introduce four complementary approaches to estimate the cumulative incidence of symptomatic COVID-19 in each state in the US as well as Puerto Rico and the District of Columbia, using a combination of excess influenza-like illness reports, COVID-19 test statistics, COVID-19 mortality reports, and a spatially structured epidemic model. Instead of relying on the estimate from a single data source or method that may be biased, we provide multiple estimates, each relying on different assumptions and data sources. Across our four approaches emerges the consistent conclusion that on April 4, 2020, the estimated case count was 5 to 50 times higher than the official positive test counts across the different states. Nationally, our estimates of COVID-19 symptomatic cases as of April 4 have a likely range of 2.3 to 4.8 million, with possibly as many as 7.6 million cases, up to 25 times greater than the cumulative confirmed cases of about 311,000. Extending our methods to May 16, 2020, we estimate that cumulative symptomatic incidence ranges from 4.9 to 10.1 million, as opposed to 1.5 million positive test counts. The proposed combination of approaches may prove useful in assessing the burden of COVID-19 during resurgences in the US and other countries with comparable surveillance systems.

Emergency preparedness for public health threats, surveillance, modelling & forecasting

In the interconnected world, safeguarding global health security is vital for maintaining public health and economic upliftment of any nation. Emergency preparedness is considered as the key to control the emerging public health challenges at both national as well as international levels. Further, the predictive information systems based on routine surveillance, disease modelling and forecasting play a pivotal role in both policy building and community participation to detect, prevent and respond to potential health threats. Therefore, reliable and timely forecasts of these untoward events could mobilize swift and effective public health responses and mitigation efforts. The present review focuses on the various aspects of emergency preparedness with special emphasis on public health surveillance, epidemiological modelling and capacity building approaches. Global coordination and capacity building, funding and commitment at the national and international levels, under the One Health framework, are crucial in combating global public health threats in a holistic manner.

COVID-19 Self-Reported Symptom Tracking Programs in the United States: Framework Synthesis

BACKGROUND

With the continued spread of COVID-19 in the United States, identifying potential outbreaks before infected individuals cross the clinical threshold is key to allowing public health officials time to ensure local health care institutions are adequately prepared. In response to this need, researchers have developed participatory surveillance technologies that allow individuals to report emerging symptoms daily so that their data can be extrapolated and disseminated to local health care authorities.

OBJECTIVE

This study uses a framework synthesis to evaluate existing self-reported symptom tracking programs in the United States for COVID-19 as an early-warning tool for probable clusters of infection. This in turn will inform decision makers and health care planners about these technologies and the usefulness of their information to aid in federal, state, and local efforts to mobilize effective current and future pandemic responses.

METHODS

Programs were identified through keyword searches and snowball sampling, then screened for inclusion. A best fit framework was constructed for all programs that met the inclusion criteria by collating information collected from each into a table for easy comparison.

RESULTS

We screened 8 programs; 6 were included in our final framework synthesis. We identified multiple common data elements, including demographic information like race, age, gender, and affiliation (all were associated with universities, medical schools, or schools of public health). Dissimilarities included collection of data regarding smoking status, mental well-being, and suspected exposure to COVID-19.

CONCLUSIONS

Several programs currently exist that track COVID-19 symptoms from participants on a semiregular basis. Coordination between symptom tracking program research teams and local and state authorities is currently lacking, presenting an opportunity for collaboration to avoid duplication of efforts and more comprehensive knowledge dissemination.

Web and phone-based COVID-19 syndromic surveillance in Canada: A cross-sectional study

BACKGROUND

Syndromic surveillance through web or phone-based polling has been used to track the course of infectious diseases worldwide. Our study objective was to describe the characteristics, symptoms, and self-reported testing rates of respondents in three different COVID-19 symptom surveys in Canada.

METHODS

This was a cross-sectional study using three distinct Canada-wide web-based surveys, and phone polling in Ontario. All three sources contained self-reported information on COVID-19 symptoms and testing. In addition to describing respondent characteristics, we examined symptom frequency and the testing rate among the symptomatic, as well as rates of symptoms and testing across respondent groups.

RESULTS

We found that over March- April 2020, 1.6% of respondents experienced a symptom on the day of their survey, 15% of Ontario households had a symptom in the previous week, and 44% of Canada-wide respondents had a symptom in the previous month. Across the three surveys, SARS-CoV-2-testing was reported in 2-9% of symptomatic responses. Women, younger and middle-aged adults (versus older adults) and Indigenous/First nations/Inuit/Métis were more likely to report at least one symptom, and visible minorities were more likely to report the combination of fever with cough or shortness of breath.

INTERPRETATION

The low rate of testing among those reporting symptoms suggests significant opportunity to expand testing among community-dwelling residents of Canada. Syndromic surveillance data can supplement public health reports and provide much-needed context to gauge the adequacy of SARS-CoV-2 testing rates.

Estimating the Cumulative Incidence of COVID-19 in the United States Using Four Complementary Approaches

Effectively designing and evaluating public health responses to the ongoing COVID-19 pandemic requires accurate estimation of the prevalence of COVID-19 across the United States (US). Equipment shortages and varying testing capabilities have however hindered the useful-ness of the official reported positive COVID-19 case counts. We introduce four complementary approaches to estimate the cumulative incidence of symptomatic COVID-19 in each state in the US as well as Puerto Rico and the District of Columbia, using a combination of excess influenza-like illness reports, COVID-19 test statistics, COVID-19 mortality reports, and a spatially structured epidemic model. Instead of relying on the estimate from a single data source or method that may be biased, we provide multiple estimates, each relying on different assumptions and data sources. Across our four approaches emerges the consistent conclusion that on April 4, 2020, the estimated case count was 5 to 50 times higher than the official positive test counts across the different states. Nationally, our estimates of COVID-19 symptomatic cases as of April 4 have a likely range of 2.2 to 4.9 million, with possibly as many as 8.1 million cases, up to 26 times greater than the cumulative confirmed cases of about 311,000. Extending our method to May 16, 2020, we estimate that cumulative symptomatic incidence ranges from 6.0 to 10.3 million, as opposed to 1.5 million positive test counts. The proposed combination of approaches may prove useful in assessing the burden of COVID-19 during resurgences in the US and other countries with comparable surveillance systems.

Surveilling Influenza Incidence With Centers for Disease Control and Prevention Web Traffic Data: Demonstration Using a Novel Dataset

Background

Influenza epidemics result in a public health and economic burden worldwide. Traditional surveillance techniques, which rely on doctor visits, provide data with a delay of 1 to 2 weeks. A means of obtaining real-time data and forecasting future outbreaks is desirable to provide more timely responses to influenza epidemics.

Objective

This study aimed to present the first implementation of a novel dataset by demonstrating its ability to supplement traditional disease surveillance at multiple spatial resolutions.

Methods

We used internet traffic data from the Centers for Disease Control and Prevention (CDC) website to determine the potential usability of this data source. We tested the traffic generated by 10 influenza-related pages in 8 states and 9 census divisions within the United States and compared it against clinical surveillance data.

Results

Our results yielded an r² value of 0.955 in the most successful case, promising results for some cases, and unsuccessful results for other cases. In the interest of scientific transparency to further the understanding of when internet data streams are an appropriate supplemental data source, we also included negative results (ie, unsuccessful models). Models that focused on a single influenza season were more successful than those that attempted to model multiple influenza seasons. Geographic resolution appeared to play a key role, with national and regional models being more successful, overall, than models at the state level.

Conclusions

These results demonstrate that internet data may be able to complement traditional influenza surveillance in some cases but not in others. Specifically, our results show that the CDC website traffic may inform national- and division-level models but not models for each individual state. In addition, our results show better agreement when the data were broken up by seasons instead of aggregated over several years. We anticipate that this work will lead to more complex nowcasting and forecasting models using this data stream.

Socioeconomic bias in influenza surveillance

Individuals in low socioeconomic brackets are considered at-risk for developing influenza-related complications and often exhibit higher than average influenza-related hospitalization rates. This disparity has been attributed to various factors, including restricted access to preventative and therapeutic health care, limited sick leave, and household structure. Adequate influenza surveillance in these at-risk populations is a critical precursor to accurate risk assessments and effective intervention. However, the United States of America's primary national influenza surveillance system (ILINet) monitors outpatient healthcare providers, which may be largely inaccessible to lower socioeconomic populations. Recent initiatives to incorporate Internet-source and hospital electronic medical records data into surveillance systems seek to improve the timeliness, coverage, and accuracy of outbreak detection and situational awareness. Here, we use a flexible statistical framework for integrating multiple surveillance data sources to evaluate the adequacy of traditional (ILINet) and next generation (BioSense 2.0 and Google Flu Trends) data for situational awareness of influenza across poverty levels. We find that ZIP Codes in the highest poverty quartile are a critical vulnerability for ILINet that the integration of next generation data fails to ameliorate.

Surveillance of COVID-19 in the General Population Using an Online Questionnaire: Report From 18,161 Respondents in China

BACKGROUND

The recent outbreak of the coronavirus disease (COVID-19) has become an international pandemic. So far, little is known about the role of an internet approach in COVID-19 participatory surveillance.

OBJECTIVE

The aim of this study is to investigate whether an online survey can provide population-level information for observing prevalence trends during the early phase of an outbreak and identifying potential risk factors of COVID-19 infection.

METHODS

A 10-item online questionnaire was developed according to medical guidelines and relevant publications. It was distributed between January 24 and February 17, 2020. The characteristics of respondents and temporal changes of various questionnaire-derived indicators were analyzed.

RESULTS

A total of 18,161 questionnaires were returned, including 6.45% (n=1171) from Wuhan City. Geographical distributions of the respondents were consistent with the population per province (R2=0.61, P<.001). History of contact significantly decreased with time, both outside Wuhan City (R2=0.35, P=.002) and outside Hubei Province (R2=0.42, P<.001). The percentage of respondents reporting a fever peaked around February 8 (R2=0.57, P<.001) and increased with a history of contact in the areas outside Wuhan City (risk ratio 1.31, 95% CI 1.13-1.52, P<.001). Male sex, advanced age, and lung diseases were associated with a higher risk of fever in the general population with a history of contact.

CONCLUSIONS

This study shows the usefulness of an online questionnaire for the surveillance of outbreaks like COVID-19 by providing information about trends of the disease and aiding the identification of potential risk factors.

Participatory Surveillance Based on Crowdsourcing During the Rio 2016 Olympic Games Using the Guardians of Health Platform: Descriptive Study

Background

With the evolution of digital media, areas such as public health are adding new platforms to complement traditional systems of epidemiological surveillance. Participatory surveillance and digital epidemiology have become innovative tools for the construction of epidemiological landscapes with citizens’ participation, improving traditional sources of information. Strategies such as these promote the timely detection of warning signs for outbreaks and epidemics in the region.

Objective

This study aims to describe the participatory surveillance platform Guardians of Health, which was used in a project conducted during the 2016 Olympic and Paralympic Games in Rio de Janeiro, Brazil, and officially used by the Brazilian Ministry of Health for the monitoring of outbreaks and epidemics.

Methods

This is a descriptive study carried out using secondary data from Guardians of Health available in a public digital repository. Based on syndromic signals, the information subsidy for decision making by policy makers and health managers becomes more dynamic and assertive. This type of information source can be used as an early route to understand the epidemiological scenario.

Results

The main result of this research was demonstrating the use of the participatory surveillance platform as an additional source of information for the epidemiological surveillance performed in Brazil during a mass gathering. The platform Guardians of Health had 7848 users who generated 12,746 reports about their health status. Among these reports, the following were identified: 161 users with diarrheal syndrome, 68 users with respiratory syndrome, and 145 users with rash syndrome.

Conclusions

It is hoped that epidemiological surveillance professionals, researchers, managers, and workers become aware of, and allow themselves to use, new tools that improve information management for decision making and knowledge production. This way, we may follow the path for a more intelligent, efficient, and pragmatic disease control system.

Online Searching and Social Media to Detect Alcohol Use Risk at Population Scale

INTRODUCTION

Harnessing engagement in online searching and social media may provide complementary information for monitoring alcohol use, informing prevention and policy evaluation, and extending knowledge available from national surveys.

METHODS

Relative search volumes for 7 alcohol-related keywords were estimated from Google Trends (data, 2014-2017), and the proportion of alcohol use-related Twitter posts (data, 2014-2015) was estimated using natural language processing. Searching/posting measures were created for all 50 U.S. states plus Washington, D.C. Survey reports of alcohol use and summaries of state alcohol policies were obtained from the Behavioral Risk Factor Surveillance System (data, 2014-2016) and the Alcohol Policy Scale. In 2018-2019, associations among searching/posting measures and same state/year Behavioral Risk Factor Surveillance System reports of recent (past-30-day) alcohol use and maximum number of drinks consumed on an occasion were estimated using logistic and linear regression, adjusting for sociodemographics and Internet use, with moderation tested in regressions that included interactions of select searching/posting measures and the Alcohol Policy Scale.

RESULTS

Recent alcohol use was reported by 52.93% of 1,297,168 Behavioral Risk Factor Surveillance System respondents, which was associated with all state-level searching/posting measures in unadjusted and adjusted models (p<0.0001). Among drinkers, most searching/posting measures were associated with maximum number of drinks consumed (p<0.0001). Associations varied with exposure to high versus low levels of state policy controls on alcohol.

CONCLUSIONS

Strong associations were found among individual alcohol use and state-level alcohol-related searching/posting measures, which were moderated by the strength of state alcohol policies. Findings support using novel personally generated data to monitor alcohol use and possibly evaluate effects of alcohol control policies.

Impact of demographic disparities in social distancing and vaccination on influenza epidemics in urban and rural regions of the United States

Background

Self-protective behaviors of social distancing and vaccination uptake vary by demographics and affect the transmission dynamics of influenza in the United States. By incorporating the socio-behavioral differences in social distancing and vaccination uptake into mathematical models of influenza transmission dynamics, we can improve our estimates of epidemic outcomes. In this study we analyze the impact of demographic disparities in social distancing and vaccination on influenza epidemics in urban and rural regions of the United States.

Methods

We conducted a survey of a nationally representative sample of US adults to collect data on their self-protective behaviors, including social distancing and vaccination to protect themselves from influenza infection. We incorporated this data in an agent-based model to simulate the transmission dynamics of influenza in the urban region of Miami Dade county in Florida and the rural region of Montgomery county in Virginia.

Results

We compare epidemic scenarios wherein the social distancing and vaccination behaviors are uniform versus non-uniform across different demographic subpopulations. We infer that a uniform compliance of social distancing and vaccination uptake among different demographic subpopulations underestimates the severity of the epidemic in comparison to differentiated compliance among different demographic subpopulations. This result holds for both urban and rural regions.

Conclusions

By taking into account the behavioral differences in social distancing and vaccination uptake among different demographic subpopulations in analysis of influenza epidemics, we provide improved estimates of epidemic outcomes that can assist in improved public health interventions for prevention and control of influenza.

Electronic supplementary material

The online version of this article (10.1186/s12879-019-3703-2) contains supplementary material, which is available to authorized users.

Causal Relationships Among Pollen Counts, Tweet Numbers, and Patient Numbers for Seasonal Allergic Rhinitis Surveillance: Retrospective Analysis

Background

Health-related social media data are increasingly used in disease-surveillance studies, which have demonstrated moderately high correlations between the number of social media posts and the number of patients. However, there is a need to understand the causal relationship between the behavior of social media users and the actual number of patients in order to increase the credibility of disease surveillance based on social media data.

Objective

This study aimed to clarify the causal relationships among pollen count, the posting behavior of social media users, and the number of patients with seasonal allergic rhinitis in the real world.

Methods

This analysis was conducted using datasets of pollen counts, tweet numbers, and numbers of patients with seasonal allergic rhinitis from Kanagawa Prefecture, Japan. We examined daily pollen counts for Japanese cedar (the major cause of seasonal allergic rhinitis in Japan) and hinoki cypress (which commonly complicates seasonal allergic rhinitis) from February 1 to May 31, 2017. The daily numbers of tweets that included the keyword “kafunshō” (or seasonal allergic rhinitis) were calculated between January 1 and May 31, 2017. Daily numbers of patients with seasonal allergic rhinitis from January 1 to May 31, 2017, were obtained from three healthcare institutes that participated in the study. The Granger causality test was used to examine the causal relationships among pollen count, tweet numbers, and the number of patients with seasonal allergic rhinitis from February to May 2017. To determine if time-variant factors affect these causal relationships, we analyzed the main seasonal allergic rhinitis phase (February to April) when Japanese cedar trees actively produce and release pollen.

Results

Increases in pollen count were found to increase the number of tweets during the overall study period (P=.04), but not the main seasonal allergic rhinitis phase (P=.05). In contrast, increases in pollen count were found to increase patient numbers in both the study period (P=.04) and the main seasonal allergic rhinitis phase (P=.01). Increases in the number of tweets increased the patient numbers during the main seasonal allergic rhinitis phase (P=.02), but not the overall study period (P=.89). Patient numbers did not affect the number of tweets in both the overall study period (P=.24) and the main seasonal allergic rhinitis phase (P=.47).

Conclusions

Understanding the causal relationships among pollen counts, tweet numbers, and numbers of patients with seasonal allergic rhinitis is an important step to increasing the credibility of surveillance systems that use social media data. Further in-depth studies are needed to identify the determinants of social media posts described in this exploratory analysis.

Feedback Between Behavioral Adaptations and Disease Dynamics

We study the feedback processes between individual behavior, disease prevalence, interventions and social networks during an influenza pandemic when a limited stockpile of antivirals is shared between the private and the public sectors. An economic model that uses prevalence-elastic demand for interventions is combined with a detailed social network and a disease propagation model to understand the feedback mechanism between epidemic dynamics, market behavior, individual perceptions, and the social network. An urban and a rural region are simulated to assess the robustness of results. Results show that an optimal split between the private and public sectors can be reached to contain the disease but the accessibility of antivirals from the private sector is skewed towards the richest income quartile. Also, larger allocations to the private sector result in wastage where individuals who do not need it are able to purchase it but who need it cannot afford it. Disease prevalence increases with household size and total contact time but not by degree in the social network, whereas wastage of antivirals decreases with degree and contact time. The best utilization of drugs is achieved when individuals with high contact time use them, who tend to be the school-aged children of large families.

A New Influenza-Tracking Smartphone App (Flu-Report) Based on a Self-Administered Questionnaire: Cross-Sectional Study

BACKGROUND

Influenza infections can spread rapidly, and influenza outbreaks are a major public health concern worldwide. Early detection of signs of an influenza pandemic is important to prevent global outbreaks. Development of information and communications technologies for influenza surveillance, including participatory surveillance systems involving lay users, has recently increased. Many of these systems can estimate influenza activity faster than the conventional influenza surveillance systems. Unfortunately, few of these influenza-tracking systems are available in Japan.

OBJECTIVE

This study aimed to evaluate the flu-tracking ability of Flu-Report, a new influenza-tracking mobile phone app that uses a self-administered questionnaire for the early detection of influenza activity.

METHODS

Flu-Report was used to collect influenza-related information (ie, dates on which influenza infections were diagnosed) from November 2016 to March 2017. Participants were adult volunteers from throughout Japan, who also provided information about their cohabiting family members. The utility of Flu-Report was evaluated by comparison with the conventional influenza surveillance information and basic information from an existing large-scale influenza-tracking system (an automatic surveillance system based on electronic records of prescription drug purchases).

RESULTS

Information was obtained through Flu-Report for approximately 10,094 volunteers. In total, 2134 participants were aged <20 years, 6958 were aged 20-59 years, and 1002 were aged ≥60 years. Between November 2016 and March 2017, 347 participants reported they had influenza or an influenza-like illness in the 2016 season. Flu-Report-derived influenza infection time series data displayed a good correlation with basic information obtained from the existing influenza surveillance system (rho, ρ=.65, P=.001). However, the influenza morbidity ratio for our participants was approximately 25% of the mean influenza morbidity ratio for the Japanese population. The Flu-Report influenza morbidity ratio was 5.06% (108/2134) among those aged <20 years, 3.16% (220/6958) among those aged 20-59 years, and 0.59% (6/1002) among those aged ≥60 years. In contrast, influenza morbidity ratios for Japanese individuals aged <20 years, 20-59 years, and ≥60 years were recently estimated at 31.97% to 37.90%, 8.16% to 9.07%, and 2.71% to 4.39%, respectively.

CONCLUSIONS

Flu-Report supports easy access to near real-time information about influenza activity via the accumulation of self-administered questionnaires. However, Flu-Report users may be influenced by selection bias, which is a common issue associated with surveillance using information and communications technologies. Despite this, Flu-Report has the potential to provide basic data that could help detect influenza outbreaks.

Participatory Disease Surveillance: Engaging Communities Directly in Reporting, Monitoring, and Responding to Health Threats

BACKGROUND

Since 2012, the International Workshop on Participatory Surveillance (IWOPS) has served as an informal network to share best practices, consult on analytic methods, and catalyze innovation to advance the burgeoning method of direct engagement of populations in voluntary monitoring of disease.

OBJECTIVE

This landscape provides an overview of participatory disease surveillance systems in the IWOPS network and orients readers to this growing field of practice.

METHODS

Authors reviewed participatory approaches that include human and animal health surveillance, both syndromic (self- reported symptoms) and event-based, and how these tools have been leveraged for disease modeling and forecasting. The authors also discuss benefits, challenges, and future directions for participatory disease surveillance.

RESULTS

There are at least 23 distinct participatory surveillance tools or programs represented in the IWOPS network across 18 countries. Organizations supporting these tools are diverse in nature.

CONCLUSIONS

Participatory disease surveillance is a promising method to complement both traditional, facility-based surveillance and newer digital epidemiology systems.