Influenza forecasting with Google Flu Trends

"What's in a name?": Using mpox as a case study to understand the importance of communication, advocacy, and information accuracy in disease nomenclature

Historically, many diseases have been named after the species or location of discovery, the discovering scientists, or the most impacted population. However, species-specific disease names often misrepresent the true reservoir; location-based disease names are frequently targeted with xenophobia; some of the discovering scientists have darker histories; and impacted populations have been stigmatized for this association. Acknowledging these concerns, the World Health Organization now proposes naming diseases after their causative pathogen or symptomatology. Recently, this guidance has been retrospectively applied to a disease at the center of an outbreak rife with stigmatization and misinformation: mpox (f.k.a. 'monkeypox'). This disease, historically endemic to west and central Africa, has prompted racist remarks as it spread globally in 2022 in an epidemic ongoing today. Moreover, its elevated prevalence among men who have sex with men has yielded increased stigma against the LGBTQ+ community. To address these prejudicial associations, 'monkeypox' was renamed 'mpox' in November 2022. We used publicly available data from Google Search Trends to determine which countries were quicker to adopt this name change-and understand factors that limit or facilitate its use. Specifically, we built regression models to quantify the relationship between 'mpox' search intensity in a given country and the country's type of political regime, robustness of sociopolitical and health systems, level of pandemic preparedness, extent of gender and educational inequalities, and temporal evolution of mpox cases through December 2023. Our results suggest that, when compared to 'monkeypox' search intensity, 'mpox' search intensity was significantly higher in countries with any history of mpox outbreaks or higher levels of LGBTQ+ acceptance; meanwhile, 'mpox' search intensity was significantly lower in countries governed by leaders who had recently propagated infectious disease misinformation. Among infectious diseases with stigmatizing names, mpox is among the first to be revised retrospectively. While the adoption of a given disease name will be context-specific-depending in part on its origins and the affected subpopulations-our study provides generalizable insights, applicable to future changes in disease nomenclature.

Predicting seasonal influenza outbreaks with regime shift-informed dynamics for improved public health preparedness

In this study, we propose a novel approach that integrates regime-shift detection with a mechanistic model to forecast the peak times of seasonal influenza. The key benefit of this approach is its ability to detect regime shifts from non-epidemic to epidemic states, which is particularly beneficial with the year-round presence of non-zero Influenza-Like Illness (ILI) data. This integration allows for the incorporation of external factors that trigger the onset of the influenza season-factors that mechanistic models alone might not adequately capture. Applied to ILI data collected in Korea from 2005 to 2020, our method demonstrated stable peak time predictions for seasonal influenza outbreaks, particularly in years characterized by unusual onset times or epidemic magnitudes.

REFORMS: Consensus-based Recommendations for Machine-learning-based Science

Machine learning (ML) methods are proliferating in scientific research. However, the adoption of these methods has been accompanied by failures of validity, reproducibility, and generalizability. These failures can hinder scientific progress, lead to false consensus around invalid claims, and undermine the credibility of ML-based science. ML methods are often applied and fail in similar ways across disciplines. Motivated by this observation, our goal is to provide clear recommendations for conducting and reporting ML-based science. Drawing from an extensive review of past literature, we present the REFORMS checklist (recommendations for machine-learning-based science). It consists of 32 questions and a paired set of guidelines. REFORMS was developed on the basis of a consensus of 19 researchers across computer science, data science, mathematics, social sciences, and biomedical sciences. REFORMS can serve as a resource for researchers when designing and implementing a study, for referees when reviewing papers, and for journals when enforcing standards for transparency and reproducibility.

Infodemiology of Influenza-like Illness: Utilizing Google Trends' Big Data for Epidemic Surveillance

Background: Influenza-like illness (ILI) encompasses symptoms similar to influenza, affecting population health. Surveillance, including Google Trends (GT), offers insights into epidemic patterns. Methods: This study used multiple regression models to analyze the correlation between ILI incidents, GT keyword searches, and climate variables during influenza outbreaks. It compared the predictive capabilities of time-series and deep learning models against ILI emergency incidents. Results: The GT searches for "fever" and "cough" were significantly associated with ILI cases (p < 0.05). Temperature had a more substantial impact on ILI incidence than humidity. Among the tested models, ARIMA provided the best predictive power. Conclusions: GT and climate data can forecast ILI trends, aiding governmental decision making. Temperature is a crucial predictor, and ARIMA models excel in forecasting ILI incidences.

Multi-step influenza forecasting through singular value decomposition and kernel ridge regression with MARCOS-guided gradient-based optimization

This research delves into the significance of influenza outbreaks in public health, particularly the importance of accurate forecasts using weekly Influenza-like illness (ILI) rates. The present work develops a novel hybrid machine-learning model by combining singular value decomposition with kernel ridge regression (SKRR). In this context, a novel hybrid model known as H-SKRR is developed by combining two robust forecasting approaches, SKRR and ridge regression, which aims to improve multi-step-ahead predictions for weekly ILI rates in Southern and Northern China. The study begins with feature selection via XGBoost in the preprocessing phase, identifying optimal precursor information guided by importance factors. It decomposes the original signal using multivariate variational mode decomposition (MVMD) to address non-stationarity and complexity. H-SKRR is implemented by incorporating significant lagged-time components across sub-components. The aggregated forecasted values from these sub-components generate ILI values for two horizons (i.e., 4-and 7-weekly ahead). Employing the gradient-based optimization (GBO) algorithm fine-tunes model parameters. Furthermore, the deep random vector functional link (dRVFL), Ridge regression, and gated recurrent unit neural network (GRU) models were employed to validate the MVMD-H-SKRR-GBO paradigm's effectiveness. The outcomes, assessed using the MARCOS (Measurement of alternatives and ranking according to compromise solution) method as a multi-criteria decision-making method, highlight the superior accuracy of the MVMD-H-SKRR-GBO model in predicting ILI rates. The results clearly highlight the exceptional performance of the MVMD-H-SKRR-GBO model, with outstanding precision demonstrated by impressive R, RMSE, IA, and U95 % values of 0.946, 0.388, 0.970, and 1.075, respectively, at t + 7.

Use of Telehealth Information for Early Detection: Insights From the COVID-19 Pandemic

Objectives. To examine whether the addition of telehealth data to existing surveillance infrastructure can improve forecasts of cases and mortality. Methods. In this observational study, we compared accuracy of 14-day forecasts using real-time data available to the National Syndromic Surveillance Program (standard forecasts) to forecasts that also included telehealth information (telehealth forecasts). The study was performed in a national telehealth service provider in 2020 serving 50 US states and the District of Columbia. Results. Among 10.5 million telemedicine encounters, 169 672 probable COVID-19 cases were diagnosed by 5050 clinicians, with a rate between 0.79 and 47.8 probable cases per 100 000 encounters per day (mean = 8.37; SD = 10.75). Publicly reported case counts ranged from 0.5 to 237 916 (mean: 53 913; SD = 47 466) and 0 to 2328 deaths (mean = 1035; SD = 550) per day. Telehealth-based forecasts improved 14-day case forecasting accuracy by 1.8 percentage points to 30.9% (P = .06) and mortality forecasting by 6.4 percentage points to 26.9% (P < .048). Conclusions. Modest improvements in forecasting can be gained from adding telehealth data to syndromic surveillance infrastructure. (Am J Public Health. 2024;114(2):218-225. https://doi.org/10.2105/AJPH.2023.307499).

Quantitatively assessing early detection strategies for mitigating COVID-19 and future pandemics

Researchers and policymakers have proposed systems to detect novel pathogens earlier than existing surveillance systems by monitoring samples from hospital patients, wastewater, and air travel, in order to mitigate future pandemics. How much benefit would such systems offer? We developed, empirically validated, and mathematically characterized a quantitative model that simulates disease spread and detection time for any given disease and detection system. We find that hospital monitoring could have detected COVID-19 in Wuhan 0.4 weeks earlier than it was actually discovered, at 2,300 cases (standard error: 76 cases) compared to 3,400 (standard error: 161 cases). Wastewater monitoring would not have accelerated COVID-19 detection in Wuhan, but provides benefit in smaller catchments and for asymptomatic or long-incubation diseases like polio or HIV/AIDS. Air travel monitoring does not accelerate outbreak detection in most scenarios we evaluated. In sum, early detection systems can substantially mitigate some future pandemics, but would not have changed the course of COVID-19.

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.

The evolution of the COVID-19 pandemic through the lens of google searches

Real-time data is essential for policymakers to adapt to a rapidly evolving situation like the COVID-19 pandemic. Using data from 221 countries and territories, we demonstrate the capacity of Google search data to anticipate reported COVID-19 cases and understand how containment policies are associated with changes in socioeconomic indicators. First, search interest in COVID-specific symptoms such as "loss of smell" strongly correlated with cases initially, but the association diminished as COVID-19 evolved; general terms such as "COVID symptoms" remained strongly associated with cases. Moreover, trends in search interest preceded trends in reported cases, particularly in the first year of the pandemic. Second, countries with more restrictive containment policies experienced greater search interest in unemployment and mental health terms after policies were implemented, indicating socio-economic externalities. Higher-income countries experienced a larger increase in searches related to unemployment and a larger reduction in relationship and family planning keywords relative to lower-income countries. The results demonstrate that real-time search interest can be a valuable tool to inform policies across multiple stages of the pandemic.

Maximum generable interest: A universal standard for Google Trends search queries

The coronavirus or COVID-19 pandemic represents a health event with far-reaching global consequences, triggering a strong search interest in related topics on the Internet worldwide. The use of search engine data has become commonplace in research, but a universal standard for comparing different works is desirable to simplify the comparison. The coronavirus pandemic's enormous impact and media coverage have triggered an exceptionally high search interest. Consequently, the maximum generable interest (MGI) on coronavirus is proposed as a universal reference for objectifying and comparing relative search interest in the future. This search interest can be explored with search engine data such as Google Trends data. Additional standards for medium and low search volumes can also be used to reflect the search interest of topics at different levels. Size standards, such as reference to MGI, may help make research more comparable and better evaluate relative search volumes. This study presents a framework for this purpose using the example of stroke.

Catch the tweet to fight the flu: Using Twitter to promote flu shots on a college campus

Objective: Over the 2018-2019 flu season we conducted a randomized controlled trial examining the efficacy of a Twitter campaign on vaccination rates. Concurrently we investigated potential interactions between digital social network structure and vaccination status. Participants: Undergratuates at a large midwestern public university were randomly assigned to an intervention (n = 353) or control (n = 349) group. Methods: Vaccination data were collected via monthly surveys. Participant Twitter data were collected through the public-facing Twitter API. Intervention impact was assessed with logistic regression. Standard network science tools examined vaccination coverage over online social networks. Results: The campaign had no effect on vaccination outcome. Receiving a flu shot the prior year had a positive impact on participant vaccination. Evidence of an interaction between digital social network structure and vaccination status was detected. Conclusions: Social media campaigns may not be sufficient for increasing vaccination rates. There may be potential for social media campaigns that leverage network structure.

Quantitatively assessing early detection strategies for mitigating COVID-19 and future pandemics

Researchers and policymakers have proposed systems to detect novel pathogens earlier than existing surveillance systems by monitoring samples from hospital patients, wastewater, and air travel, in order to mitigate future pandemics. How much benefit would such systems offer? We developed, empirically validated, and mathematically characterized a quantitative model that simulates disease spread and detection time for any given disease and detection system. We find that hospital monitoring could have detected COVID-19 in Wuhan 0.4 weeks earlier than it was actually discovered, at 2,300 cases (standard error: 76 cases) compared to 3,400 (standard error: 161 cases). Wastewater monitoring would not have accelerated COVID-19 detection in Wuhan, but provides benefit in smaller catchments and for asymptomatic or long-incubation diseases like polio or HIV/AIDS. Monitoring of air travel provides little benefit in most scenarios we evaluated. In sum, early detection systems can substantially mitigate some future pandemics, but would not have changed the course of COVID-19.

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.

Web-based surveillance of respiratory infection outbreaks: retrospective analysis of Italian COVID-19 epidemic waves using Google Trends

Introduction

Large-scale diagnostic testing has been proven insufficient to promptly monitor the spread of the Coronavirus disease 2019. Electronic resources may provide better insight into the early detection of epidemics. We aimed to retrospectively explore whether the Google search volume has been useful in detecting Severe Acute Respiratory Syndrome Coronavirus outbreaks early compared to the swab-based surveillance system.

Methods

The Google Trends website was used by applying the research to three Italian regions (Lombardy, Marche, and Sicily), covering 16 million Italian citizens. An autoregressive-moving-average model was fitted, and residual charts were plotted to detect outliers in weekly searches of five keywords. Signals that occurred during periods labelled as free from epidemics were used to measure Positive Predictive Values and False Negative Rates in anticipating the epidemic wave occurrence.

Results

Signals from "fever," "cough," and "sore throat" showed better performance than those from "loss of smell" and "loss of taste." More than 80% of true epidemic waves were detected early by the occurrence of at least an outlier signal in Lombardy, although this implies a 20% false alarm signals. Performance was poorer for Sicily and Marche.

Conclusion

Monitoring the volume of Google searches can be a valuable tool for early detection of respiratory infectious disease outbreaks, particularly in areas with high access to home internet. The inclusion of web-based syndromic keywords is promising as it could facilitate the containment of COVID-19 and perhaps other unknown infectious diseases in the future.

Joint COVID-19 and influenza-like illness forecasts in the United States using internet search information

BACKGROUND

As the prolonged COVID-19 pandemic continues, severe seasonal Influenza (flu) may happen alongside COVID-19. This could cause a "twindemic", in which there are additional burdens on health care resources and public safety compared to those occurring in the presence of a single infection. Amidst the raising trend of co-infections of the two diseases, forecasting both Influenza-like Illness (ILI) outbreaks and COVID-19 waves in a reliable and timely manner becomes more urgent than ever. Accurate and real-time joint prediction of the twindemic aids public health organizations and policymakers in adequate preparation and decision making. However, in the current pandemic, existing ILI and COVID-19 forecasting models face shortcomings under complex inter-disease dynamics, particularly due to the similarities in symptoms and healthcare-seeking patterns of the two diseases.

METHODS

Inspired by the interconnection between ILI and COVID-19 activities, we combine related internet search and bi-disease time series information for the U.S. national level and state level forecasts. Our proposed ARGOX-Joint-Ensemble adopts a new ensemble framework that integrates ILI and COVID-19 disease forecasting models to pool the information between the two diseases and provide joint multi-resolution and multi-target predictions. Through a winner-takes-all ensemble fashion, our framework is able to adaptively select the most predictive COVID-19 or ILI signals.

RESULTS

In the retrospective evaluation, our model steadily outperforms alternative benchmark methods, and remains competitive with other publicly available models in both point estimates and probabilistic predictions (including intervals).

CONCLUSIONS

The success of our approach illustrates that pooling information between the ILI and COVID-19 leads to improved forecasting models than individual models for either of the disease.

Association Between Internet Use and Physical Health, Mental Health, and Subjective Health in Middle-aged and Older Adults: Nationally Representative Cross-sectional Survey in China

BACKGROUND

Internet use is an important means of accessing health-related information. Identifying the associations between internet use and health outcomes could provide insight into strategies for improving public health among middle-aged and older adults (45 years and up).

OBJECTIVE

This study aimed to examine the relationship between internet use and health outcomes in middle-aged and older adults.

METHODS

Data were obtained from the 2018 China Health and Retirement Longitudinal Study. Physical, mental, and subjective health were assessed using the Activities of Daily Living (ADL) Scale, the 10-item Center for Epidemiologic Studies Depression Scale, and the 3-level Self-Rated Health Scale, respectively. The chi-square test and rank sum test were used to explore whether internet use was associated with health status. A multivariate logistic regression model was used to determine this association further after controlling for the confounding factors.

RESULTS

Overall, 13% (1752/13,474) of the participants used the internet. Regression analyses revealed that the prevalence of depression (odds ratio [OR] 0.59, 95% CI 0.52-0.68; P<.001), negative self-rated health (OR 0.68, 95% CI 0.61-0.76; P<.001), and difficulty with ADL (OR 0.48, 95% CI 0.39-0.60; P<.001) in the participating middle-aged and older adult was lower in those using the internet than nonusers. After controlling for confounding factors, internet use was found to be negatively associated with difficulty with ADL (urban: OR 0.44, 95% CI 0.32-0.61; P<.001 vs rural: OR 0.55, 95% CI 0.41-0.75; P<.001), depression (urban: OR 0.69, 95% CI 0.57-0.84; P<.001 vs rural: OR 0.52, 95% CI: 0.43-0.63; P<.001), and self-rated health status (urban: OR 0.70, 95% CI 0.61-0.81; P<.001 vs rural: OR 0.67, 95% CI 0.57-0.78; P<.001) among middle-aged and older adults in both urban and rural areas.

CONCLUSIONS

Internet use had a positive effect on the physical and mental health of middle-aged and older adults who participated in this study. However, the internet usage rate remains low among older Chinese people. Therefore, the internet penetration rate should be a priority.

Internet searches for "insomnia" and "suicide" mediated by stay-at-home behaviors in 45 countries during the first 12 months of the COVID-19 pandemic

BACKGROUND

This study aimed to investigate (1) the mental health impacts (i.e., insomnia and suicide ideas) of the COVID-19 pandemic and (2) the mediation effects of stay-at-home levels on those impacts.

METHODS

This study investigated monthly national COVID-19 deaths, stay-at-home levels, and internet searches for words for "insomnia" and "suicide" across 45 countries during the first year of the COVID-19 pandemic (March 1, 2020, to February 28, 2021). We used the changes of internet search volumes for "insomnia" and "suicide" (from the Google Trends database) to represent the mental health impacts, and the time of cell phone activity at the residence (from Google Location History) to estimate the stay-at-home effects. We computed the proportion mediated (PM) caused by stay-at-home levels in the COVID-19 impacts on insomnia and suicide ideas, respectively.

RESULTS

Throughout the first year of the COVID-19 pandemic, national COVID-19 deaths significantly correlated to increased internet searches for "insomnia" but decreased searches for "suicide". In addition, the mediation effect was significant in the first six-month of COVID-19-related increases in insomnia (PM = 42.6 %, p = 0.016), but this effect was not significant (PM = 13.1 %, p = 0.270) in the second six-month. By contrast, the mediation effect was not significant in the first six-month of COVID-19-related decrease in suicide ideation (PM = 8.1 %, p = 0.180), but this effect was significant (PM = 39.6 %, p = 0.014) in the second six-month.

CONCLUSIONS

Stay-at-home levels significantly mediated both increased insomnia and decreased suicide ideas, but within different time frames.

Development of a Machine Learning Model to Estimate US Firearm Homicides in Near Real Time

Importance

Firearm homicides are a major public health concern; lack of timely mortality data presents considerable challenges to effective response. Near real-time data sources offer potential for more timely estimation of firearm homicides.

Objective

To estimate near real-time burden of weekly and annual firearm homicides in the US.

Design, Setting, and Participants

In this prognostic study, anonymous, longitudinal time series data were obtained from multiple data sources, including Google and YouTube search trends related to firearms (2014-2019), emergency department visits for firearm injuries (National Syndromic Surveillance Program, 2014-2019), emergency medical service activations for firearm-related injuries (biospatial, 2014-2019), and National Domestic Violence Hotline contacts flagged with the keyword firearm (2016-2019). Data analysis was performed from September 2021 to September 2022.

Main Outcomes and Measures

Weekly estimates of US firearm homicides were calculated using a 2-phase pipeline, first fitting optimal machine learning models for each data stream and then combining the best individual models into a stacked ensemble model. Model accuracy was assessed by comparing predictions of firearm homicides in 2019 to actual firearm homicides identified by National Vital Statistics System death certificates. Results were also compared with a SARIMA (seasonal autoregressive integrated moving average) model, a common method to forecast injury mortality.

Results

Both individual and ensemble models yielded highly accurate estimates of firearm homicides. Individual models' mean error for weekly estimates of firearm homicides (root mean square error) varied from 24.95 for emergency department visits to 31.29 for SARIMA forecasting. Ensemble models combining data sources had lower weekly mean error and higher annual accuracy than individual data sources: the all-source ensemble model had a weekly root mean square error of 24.46 deaths and full-year accuracy of 99.74%, predicting the total number of firearm homicides in 2019 within 38 deaths for the entire year (compared with 95.48% accuracy and 652 deaths for the SARIMA model). The model decreased the time lag of reporting weekly firearm homicides from 7 to 8 months to approximately 6 weeks.

Conclusions and Relevance

In this prognostic study of diverse secondary data on machine learning, ensemble modeling produced accurate near real-time estimates of weekly and annual firearm homicides and substantially decreased data source time lags. Ensemble model forecasts can accelerate public health practitioners' and policy makers' ability to respond to unanticipated shifts in firearm homicides.

Surveillance of communicable diseases using social media: A systematic review

BACKGROUND

Communicable diseases pose a severe threat to public health and economic growth. The traditional methods that are used for public health surveillance, however, involve many drawbacks, such as being labor intensive to operate and resulting in a lag between data collection and reporting. To effectively address the limitations of these traditional methods and to mitigate the adverse effects of these diseases, a proactive and real-time public health surveillance system is needed. Previous studies have indicated the usefulness of performing text mining on social media.

OBJECTIVE

To conduct a systematic review of the literature that used textual content published to social media for the purpose of the surveillance and prediction of communicable diseases.

METHODOLOGY

Broad search queries were formulated and performed in four databases. Both journal articles and conference materials were included. The quality of the studies, operationalized as reliability and validity, was assessed. This qualitative systematic review was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

RESULTS

Twenty-three publications were included in this systematic review. All studies reported positive results for using textual social media content to surveille communicable diseases. Most studies used Twitter as a source for these data. Influenza was studied most frequently, while other communicable diseases received far less attention. Journal articles had a higher quality (reliability and validity) than conference papers. However, studies often failed to provide important information about procedures and implementation.

CONCLUSION

Text mining of health-related content published on social media can serve as a novel and powerful tool for the automated, real-time, and remote monitoring of public health and for the surveillance and prediction of communicable diseases in particular. This tool can address limitations related to traditional surveillance methods, and it has the potential to supplement traditional methods for public health surveillance.

Harnessing Google Health Trends Data for Epidemiologic Research

Interest in using internet search data, such as that from the Google Health Trends Application Programming Interface (GHT-API), to measure epidemiologically relevant exposures or health outcomes is growing due to their accessibility and timeliness. Researchers enter search term(s), geography, and time period, and the GHT-API returns a scaled probability of that search term, given all searches within the specified geographic-time period. In this study, we detailed a method for using these data to measure a construct of interest in 5 iterative steps: first, identify phrases the target population may use to search for the construct of interest; second, refine candidate search phrases with incognito Google searches to improve sensitivity and specificity; third, craft the GHT-API search term(s) by combining the refined phrases; fourth, test search volume and choose geographic and temporal scales; and fifth, retrieve and average multiple samples to stabilize estimates and address missingness. An optional sixth step involves accounting for changes in total search volume by normalizing. We present a case study examining weekly state-level child abuse searches in the United States during the coronavirus disease 2019 pandemic (January 2018 to August 2020) as an application of this method and describe limitations.

Long Short-term Memory-Based Prediction of the Spread of Influenza-Like Illness Leveraging Surveillance, Weather, and Twitter Data: Model Development and Validation

BACKGROUND

The potential to harness the plurality of available data in real time along with advanced data analytics for the accurate prediction of influenza-like illness (ILI) outbreaks has gained significant scientific interest. Different methodologies based on the use of machine learning techniques and traditional and alternative data sources, such as ILI surveillance reports, weather reports, search engine queries, and social media, have been explored with the ultimate goal of being used in the development of electronic surveillance systems that could complement existing monitoring resources.

OBJECTIVE

The scope of this study was to investigate for the first time the combined use of ILI surveillance data, weather data, and Twitter data along with deep learning techniques toward the development of prediction models able to nowcast and forecast weekly ILI cases. By assessing the predictive power of both traditional and alternative data sources on the use case of ILI, this study aimed to provide a novel approach for corroborating evidence and enhancing accuracy and reliability in the surveillance of infectious diseases.

METHODS

The model's input space consisted of information related to weekly ILI surveillance, web-based social (eg, Twitter) behavior, and weather conditions. For the design and development of the model, relevant data corresponding to the period of 2010 to 2019 and focusing on the Greek population and weather were collected. Long short-term memory (LSTM) neural networks were leveraged to efficiently handle the sequential and nonlinear nature of the multitude of collected data. The 3 data categories were first used separately for training 3 LSTM-based primary models. Subsequently, different transfer learning (TL) approaches were explored with the aim of creating various feature spaces combining the features extracted from the corresponding primary models' LSTM layers for the latter to feed a dense layer.

RESULTS

The primary model that learned from weather data yielded better forecast accuracy (root mean square error [RMSE]=0.144; Pearson correlation coefficient [PCC]=0.801) than the model trained with ILI historical data (RMSE=0.159; PCC=0.794). The best performance was achieved by the TL-based model leveraging the combination of the 3 data categories (RMSE=0.128; PCC=0.822).

CONCLUSIONS

The superiority of the TL-based model, which considers Twitter data, weather data, and ILI surveillance data, reflects the potential of alternative public sources to enhance accurate and reliable prediction of ILI spread. Despite its focus on the use case of Greece, the proposed approach can be generalized to other locations, populations, and social media platforms to support the surveillance of infectious diseases with the ultimate goal of reinforcing preparedness for future epidemics.

Influenza surveillance with Baidu index and attention-based long short-term memory model

BACKGROUND

The prediction and prevention of influenza is a public health issue of great concern, and the study of timely acquisition of influenza transmission trend has become an important research topic. For achieving more quicker and accurate detection and prediction, the data recorded on the Internet, especially on the search engine from Google or Baidu are widely introduced into this field. Moreover, with the development of intelligent technology and machine learning algorithm, many updated and advanced trend tracking and forecasting methods are also being used in this research problem.

METHODS

In this paper, a new recurrent neural network architecture, attention-based long short-term memory model is proposed for influenza surveillance. This is a kind of deep learning model which is trained by processing from Baidu Index series so as to fit the real influenza survey time series. Previous studies on influenza surveillance by Baidu Index mostly used traditional autoregressive moving average model or classical machine learning models such as logarithmic linear regression, support vector regression or multi-layer perception model to fit influenza like illness data, which less considered the deep learning structure. Meanwhile, some new model that considered the deep learning structure did not take into account the application of Baidu index data. This study considers introducing the recurrent neural network with long short-term memory combined with attention mechanism into the influenza surveillance research model, which not only fits the research problems well in model structure, but also provides research methods based on Baidu index.

RESULTS

The actual survey data and Baidu Index data are used to train and test the proposed attention-based long short-term memory model and the other comparison models, so as to iterate the value of the model parameters, and to describe and predict the influenza epidemic situation. The experimental results show that our proposed model has better performance in the mean absolute error, mean absolute percentage error, index of agreement and other indicators than the other comparison models.

CONCLUSION

Our proposed attention-based long short-term memory model vividly verifies the ability of this attention-based long short-term memory structure for better surveillance and prediction the trend of influenza. In comparison with some of the latest models and methods in this research field, the model we proposed is also excellent in effect, even more lightweight and robust. Future research direction can consider fusing multimodal data based on this model and developing more application scenarios.

Using digital traces to build prospective and real-time county-level early warning systems to anticipate COVID-19 outbreaks in the United States

Coronavirus disease 2019 (COVID-19) continues to affect the world, and the design of strategies to curb disease outbreaks requires close monitoring of their trajectories. We present machine learning methods that leverage internet-based digital traces to anticipate sharp increases in COVID-19 activity in U.S. counties. In a complementary direction to the efforts led by the Centers for Disease Control and Prevention (CDC), our models are designed to detect the time when an uptrend in COVID-19 activity will occur. Motivated by the need for finer spatial resolution epidemiological insights, we build upon previous efforts conceived at the state level. Our methods-tested in an out-of-sample manner, as events were unfolding, in 97 counties representative of multiple population sizes across the United States-frequently anticipated increases in COVID-19 activity 1 to 6 weeks before local outbreaks, defined when the effective reproduction number R_t becomes larger than 1 for a period of 2 weeks.

Short-term local predictions of COVID-19 in the United Kingdom using dynamic supervised machine learning algorithms

Background

Short-term prediction of COVID-19 epidemics is crucial to decision making. We aimed to develop supervised machine-learning algorithms on multiple digital metrics including symptom search trends, population mobility, and vaccination coverage to predict local-level COVID-19 growth rates in the UK.

Methods

Using dynamic supervised machine-learning algorithms based on log-linear regression, we explored optimal models for 1-week, 2-week, and 3-week ahead prediction of COVID-19 growth rate at lower tier local authority level over time. Model performance was assessed by calculating mean squared error (MSE) of prospective prediction, and naïve model and fixed-predictors model were used as reference models. We assessed real-time model performance for eight five-weeks-apart checkpoints between 1st March and 14th November 2021. We developed an online application (COVIDPredLTLA) that visualised the real-time predictions for the present week, and the next one and two weeks.

Results

Here we show that the median MSEs of the optimal models for 1-week, 2-week, and 3-week ahead prediction are 0.12 (IQR: 0.08–0.22), 0.29 (0.19–0.38), and 0.37 (0.25–0.47), respectively. Compared with naïve models, the optimal models maintain increased accuracy (reducing MSE by a range of 21–35%), including May–June 2021 when the delta variant spread across the UK. Compared with the fixed-predictors model, the advantage of dynamic models is observed after several iterations of update.

Conclusions

With flexible data-driven predictors selection process, our dynamic modelling framework shows promises in predicting short-term changes in COVID-19 cases. The online application (COVIDPredLTLA) could assist decision-making for control measures and planning of healthcare capacity in future epidemic growths.

Understanding how COVID-19 case numbers will change in the short term in local contexts is helpful for planning public health measures. Here, using population data on Google searches on COVID-19 symptoms, mobility (i.e., movement of people between locations), and vaccination coverage, we develop statistical methods that forecast changes in COVID-19 case numbers for the next one to three weeks in different local authorities in the UK. We further develop an online application that provides visualisations of these forecasts. We show that our programmes achieve better predictive accuracy compared with two existing models. This highlights the promise of our methods in forecasting future local COVID-19 outbreaks.

Wang et al. predict local COVID-19 incidence one to three weeks ahead based on symptom search trends, population mobility, and vaccination coverage. Their dynamic model is more accurate than naïve and fixed-predictor models.

Early Detection of COVID-19 outbreaks using Textual Analysis of Electronic Medical Records

Purpose

Our objective was to develop a tool promoting early detection of COVID-19 cases by focusing epidemiological investigations and PCR examinations during a period of limited testing capabilities.

Methods

We developed an algorithm for analyzing medical records recorded by healthcare providers in the Israeli Defense Forces. The algorithm utilized textual analysis to detect patients presenting with suspicious symptoms and was tested among 92 randomly selected units. Detection of a potential cluster of patients in a unit prompted a focused epidemiological investigation aided by data provided by the algorithm.

Results

During a month of follow up, the algorithm has flagged 17 of the units for investigation. The subsequent epidemiological investigations led to the testing of 78 persons and the detection of eight cases in four clusters that were previously gone unnoticed. The resulting positive test rate of 10.25% was five time higher than the IDF average at the time of the study. No cases of COVID-19 in the examined units were missed by the algorithm.

Conclusions

This study depicts the successful development and large scale deployment of a textual analysis based algorithm for early detection of COVID-19 cases, demonstrating the potential of natural language processing of medical text as a tool for promoting public health.

Utilizing Internet Search Volume to Monitor Stages of Change in Vaccine Hesitancy During the COVID-19 Outbreaks

Real-time vaccine hesitancy surveillance is needed to better understand changes in vaccination behaviors. We aim to understand the association between coronavirus disease 2019 (COVID-19) outbreaks and population vaccine hesitancy and to monitor the dynamic changes in vaccination behaviors. We used the autoregressive integrated moving average model to examine the association between daily internet search volume for vaccines and two waves of COVID-19 local outbreaks in Taiwan from 19 March to 25 May, 2021. During the small-scale outbreak, the search volume increased significantly for 7 out of 22 days with an average increase of 17.3% ± 10.7% from the expected search volume. During the large-scale outbreak, the search volume increased significantly for 14 out of 14 days, with an average increase of 58.4% ± 14.7%. There was a high correlation between the search volume and the number of domestic cases (r = 0.71, P < 0.001). Google Trends serves as a timely indicator to monitor the extent of population vaccine willingness.

Introduction to supervised machine learning in clinical epidemiology

Machine learning refers to a series of processes in which a computer finds rules from a vast amount of data. With recent advances in computer technology and the availability of a wide variety of health data, machine learning has rapidly developed and been applied in medical research. Currently, there are three types of machine learning: supervised, unsupervised, and reinforcement learning. In medical research, supervised learning is commonly used for diagnoses and prognoses, while unsupervised learning is used for phenotyping a disease, and reinforcement learning for maximizing favorable results, such as optimization of total patients' waiting time in the emergency department. The present article focuses on the concept and application of supervised learning in medicine, the most commonly used machine learning approach in medicine, and provides a brief explanation of four algorithms widely used for prediction (random forests, gradient-boosted decision tree, support vector machine, and neural network). Among these algorithms, the neural network has further developed into deep learning algorithms to solve more complex tasks. Along with simple classification problems, deep learning is commonly used to process medical imaging, such as retinal fundus photographs for diabetic retinopathy diagnosis. Although machine learning can bring new insights into medicine by processing a vast amount of data that are often beyond human capacity, algorithms can also fail when domain knowledge is neglected. The combination of algorithms and human cognitive ability is a key to the successful application of machine learning in medicine.

Addressing delayed case reporting in infectious disease forecast modeling

Infectious disease forecasting is of great interest to the public health community and policymakers, since forecasts can provide insight into disease dynamics in the near future and inform interventions. Due to delays in case reporting, however, forecasting models may often underestimate the current and future disease burden.

In this paper, we propose a general framework for addressing reporting delay in disease forecasting efforts with the goal of improving forecasts. We propose strategies for leveraging either historical data on case reporting or external internet-based data to estimate the amount of reporting error. We then describe several approaches for adapting general forecasting pipelines to account for under- or over-reporting of cases. We apply these methods to address reporting delay in data on dengue fever cases in Puerto Rico from 1990 to 2009 and to reports of influenza-like illness (ILI) in the United States between 2010 and 2019. Through a simulation study, we compare method performance and evaluate robustness to assumption violations. Our results show that forecasting accuracy and prediction coverage almost always increase when correction methods are implemented to address reporting delay. Some of these methods required knowledge about the reporting error or high quality external data, which may not always be available. Provided alternatives include excluding recently-reported data and performing sensitivity analysis. This work provides intuition and guidance for handling delay in disease case reporting and may serve as a useful resource to inform practical infectious disease forecasting efforts.

The public health community and policymakers are interested in using models to predict future disease rates using information about disease rates in the past. However, our data about the recent past are less reliable than older data, due to a time lag between someone getting sick and their subsequent diagnosis being officially reported. In this paper, we describe strategies to correct reported disease rates from the recent past to account for disease diagnoses that haven’t yet been reported. Using more accurate information about the recent past, we can do a better job predicting what will happen in the future.

Internet searches and heat-related emergency department visits in the United States

Emerging research suggests that internet search patterns may provide timely, actionable insights into adverse health impacts from, and behavioral responses to, days of extreme heat, but few studies have evaluated this hypothesis, and none have done so across the United States. We used two-stage distributed lag nonlinear models to quantify the interrelationships between daily maximum ambient temperature, internet search activity as measured by Google Trends, and heat-related emergency department (ED) visits among adults with commercial health insurance in 30 US metropolitan areas during the warm seasons (May to September) from 2016 to 2019. Maximum daily temperature was positively associated with internet searches relevant to heat, and searches were in turn positively associated with heat-related ED visits. Moreover, models combining internet search activity and temperature had better predictive ability for heat-related ED visits compared to models with temperature alone. These results suggest that internet search patterns may be useful as a leading indicator of heat-related illness or stress.

The supply and demand of news during COVID-19 and assessment of questionable sources production

Misinformation threatens our societies, but little is known about how the production of news by unreliable sources relates to supply and demand dynamics. We exploit the burst of news production triggered by the COVID-19 outbreak through an Italian database partially annotated for questionable sources. We compare news supply with news demand, as captured by Google Trends data. We identify the Granger causal relationships between supply and demand for the most searched keywords, quantifying the inertial behaviour of the news supply. Focusing on COVID-19 news, we find that questionable sources are more sensitive than general news production to people's interests, especially when news supply and demand mismatched. We introduce an index assessing the level of questionable news production solely based on the available volumes of news and searches. We contend that these results can be a powerful asset in informing campaigns against disinformation and providing news outlets and institutions with potentially relevant strategies.

Query-based-learning mortality-related decoders for the developed island economy

Search volumes from Google Trends over clear-defined temporal and spatial scales were reported beneficial in predicting influenza or disease outbreak. Recent studies showed Wiener Model shares merits of interpretability, implementation, and adaptation to nonlinear fluctuation in terms of real-time decoding. Previous work reported Google Trends effectively predicts death-related trends for the continent economy, yet whether it applies to the island economy is unclear. To this end, a framework of the mortality-related model for a developed island economy Taiwan was built based on potential death causes from Google Trends, aiming to provide new insights into death-related online search behavior at a population level. Our results showed estimated trends based on the Wiener model significantly correlated to actual trends, outperformed those with multiple linear regression and seasonal autoregressive integrated moving average. Meanwhile, apart from that involved all possible features, two other sets of feature selecting strategies were proposed to optimize pre-trained models, either by weights or waveform periodicity of features, resulting in estimated death-related dynamics along with spectrums of risk factors. In general, high-weight features were beneficial to both "die" and "death", whereas features that possessed clear periodic patterns contributed more to "death". Of note, normalization before modeling improved decoding performances.

Learning to sense from events via semantic variational autoencoder

In this paper, we introduce the concept of learning to sense, which aims to emulate a complex characteristic of human reasoning: the ability to monitor and understand a set of interdependent events for decision-making processes. Event datasets are composed of textual data and spatio-temporal features that determine where and when a given phenomenon occurred. In learning to sense, related events are mapped closely to each other in a semantic vector space, thereby identifying that they contain similar contextual meaning. However, learning a semantic vector space that satisfies both textual similarities and spatio-temporal constraints is a crucial challenge for event analysis and sensing. This paper investigates a Semantic Variational Autoencoder (SVAE) to fine-tune pre-trained embeddings according to both textual and spatio-temporal events of the class of interest. Experiments involving more than one hundred sensors show that our SVAE outperforms a competitive one-class classification baseline. Moreover, our proposal provides desirable learning requirements to sense scenarios, such as visualization of the sensor decision function and heat maps with the sensor’s geographic impact.

Anosmia-related internet search and the course of the first wave of the COVID-19 pandemic in the United States

Background

The current pandemic of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) was first reported in Wuhan, China. Although the first case in the United States was reported on Jan 20, 2020 in Washington, the early pandemic time course is uncertain. One approach with the potential to provide more insight into this time course is the examination of search activity. This study analyzed US search data prior to the first press release of anosmia as an early symptom (March 20, 2020).

Methods

Daily internet search query data was obtained from Google Trends (September 20th to March 20th for 2015 to 2020) both for the United States and on a state-by-state basis. Normalized anosmia-related search activity for the years prior to the pandemic was averaged to obtain a baseline level. Cross-correlations were performed to determine the time-lag between changes in search activity and SARS-CoV-2 cases/deaths.

Results

Only New York showed both significant increases in anosmia-related terms during the pandemic year as well as a significant lag (6 days) between increases in search activity and the number of cases/deaths attributed to SARS-CoV-2.

Conclusions

There is no evidence from search activity to suggest earlier spread of SARS-CoV-2 than has been previously reported. The increase in anosmia-related searches preceded increases in SARS-CoV-2 cases/deaths by 6 days, but this was only significant over the background noise of searches for other reasons in the setting of a very large outbreak (New York in the spring of 2020).

•

Recent work has suggested using digital epidemiology to follow pandemics.

•

In our view, these previous studies have several methodological errors.

•

They used correlations long after anosmia symptoms were well documented in the media.

•

We demonstrated significant issues with digital surveillance during such a high interest event.

•

A large signal is required to overcome noise introduced by searches for other reasons.

Examining the Correlation of Google Influenza Trend with Hospital Data: Retrospective Study

Introduction

Many studies have explored social media and users search activities such as Google Trends to predict and detect influenza activities. Studies that examined Google Trends correlation with the actual hospital influenza cases were conducted in non-tropical regions that have clearly defined seasons. Tropical areas are known for having less-defined seasonality and the extent of Google Trends concordance with actual influenza cases is unknown for these areas. The goal of this study is to compare Google Trends with hospital cases in tropical regions.

Methods

We analyzed 48,263 influenza cases in the time period of 2010 to 2019. The cases were retrieved from central hospital medical records in tropical regions using the corresponding codes for influenza ICD-10 AM. Cases from the medical records were compared with Google Trends to determine trends, seasonality, and correlation.

Results

Graphically, there were some similar areas of the trend, but cross-correlation analysis did not show any significant correlation between hospital and Google Trends with a maximum correlation rate of 0.300. Seasonality analysis showed a clear pattern that peaked around November in Google Trends while hospital data showed less defined seasonality with a smaller peak occurring at the end of December and beginning of January.

Conclusion

Based on the results, there is a weak correlation between Google Trends and hospital data. More innovative methods are emerging to predict influenza activity using social media and user search data and further study is needed to examine the concurrent trends derived using these methods across regions that have different humidity levels and temperatures.

Retrospective analysis and time series forecasting with automated machine learning of ascariasis, enterobiasis and cystic echinococcosis in Romania

The epidemiology of neglected tropical diseases (NTD) is persistently underprioritized, despite NTD being widespread among the poorest populations and in the least developed countries on earth. This situation necessitates thorough and efficient public health intervention. Romania is at the brink of becoming a developed country. However, this South-Eastern European country appears to be a region that is susceptible to an underestimated burden of parasitic diseases despite recent public health reforms. Moreover, there is an evident lack of new epidemiologic data on NTD after Romania's accession to the European Union (EU) in 2007. Using the national ICD-10 dataset for hospitalized patients in Romania, we generated time series datasets for 2008-2018. The objective was to gain deep understanding of the epidemiological distribution of three selected and highly endemic parasitic diseases, namely, ascariasis, enterobiasis and cystic echinococcosis (CE), during this period and forecast their courses for the ensuing two years. Through descriptive and inferential analysis, we observed a decline in case numbers for all three NTD. Several distributional particularities at regional level emerged. Furthermore, we performed predictions using a novel automated time series (AutoTS) machine learning tool and could interestingly show a stable course for these parasitic NTD. Such predictions can help public health officials and medical organizations to implement targeted disease prevention and control. To our knowledge, this is the first study involving a retrospective analysis of ascariasis, enterobiasis and CE on a nationwide scale in Romania. It is also the first to use AutoTS technology for parasitic NTD.

A general method for estimating the prevalence of influenza-like-symptoms with Wikipedia data

Influenza is an acute respiratory seasonal disease that affects millions of people worldwide and causes thousands of deaths in Europe alone. Estimating in a fast and reliable way the impact of an illness on a given country is essential to plan and organize effective countermeasures, which is now possible by leveraging unconventional data sources like web searches and visits. In this study, we show the feasibility of exploiting machine learning models and information about Wikipedia’s page views of a selected group of articles to obtain accurate estimates of influenza-like illnesses incidence in four European countries: Italy, Germany, Belgium, and the Netherlands. We propose a novel language-agnostic method, based on two algorithms, Personalized PageRank and CycleRank, to automatically select the most relevant Wikipedia pages to be monitored without the need for expert supervision. We then show how our model can reach state-of-the-art results by comparing it with previous solutions.

Improving Google Flu Trends for COVID-19 estimates using Weibo posts

While incomplete non-medical data has been integrated into prediction models for epidemics, the accuracy and the generalizability of the data are difficult to guarantee. To comprehensively evaluate the ability and applicability of using social media data to predict the development of COVID-19, a new confirmed case prediction algorithm improving the Google Flu Trends algorithm is established, called Weibo COVID-19 Trends (WCT), based on the post dataset generated by all users in Wuhan on Sina Weibo. A genetic algorithm is designed to select the keyword set for filtering COVID-19 related posts. WCT can constantly outperform the highest average test score in the training set between daily new confirmed case counts and the prediction results. It remains to produce the best prediction results among other algorithms when the number of forecast days increases from one to eight days with the highest correlation score from 0.98 (P ​< 0.01) to 0.86 (P ​< 0.01) during all analysis period. Additionally, WCT effectively improves the Google Flu Trends algorithm's shortcoming of overestimating the epidemic peak value. This study offers a highly adaptive approach for feature engineering of third-party data in epidemic prediction, providing useful insights for the prediction of newly emerging infectious diseases at an early stage.

The Impact of the Integrated Development of AI and Energy Industry on Regional Energy Industry: A Case of China

With the advent of the Energy 4.0 era, the adoption of “Internet + artificial intelligence” systems will enable the transformation and upgrading of the traditional energy industry. This will alleviate the energy and environmental problems that China is currently facing. The integrated development of artificial intelligence and the energy industry has become inevitable in the development of future energy systems. This study applied a comprehensive evaluation index to the energy industry to calculate the comprehensive development index of the energy industry in 30 provinces of China from 2000 to 2017. Then, taking Guangdong and Jiangsu as examples, the synthetic control method was used to explore the direction and intensity of the integrated development of artificial intelligence and the energy industry on the comprehensive development level of the local energy industry. The results showed that when artificial intelligence (AI) and the energy industry achieved a stable coupled development without the need to move to the coordination stage, the coupling effect promoted the development of the regional energy industry, and the annual growth rate of the comprehensive development index was above 20%. This coupling effect passed the placebo test and ranking test and was significant at the 10% level, indicating the robustness and validity of the experimental results, which strongly confirmed the great potential of AI in re-empowering traditional industries from the data perspective. Based on the findings, corresponding policy recommendations were proposed on how to promote the development of inter-regional AI, how the government, enterprises, and universities could cooperate to promote the coordinated development of AI and energy, and how to guide the integration process of regional AI and energy industries according to local conditions, in order to maximize the technological dividend of AI and help the construction of smart energy in China.

Google trends in "anatomy": pre-pandemic versus during COVID-19 pandemic

Purpose

The purpose of the present study was to assess the online search behavior for the keyword “anatomy” worldwide and to compare the pre-pandemic and during COVID-19 pandemic scenario for the same.

Methods

Google trends tool was used for the assessment of the search behavior for the term “anatomy”. The data, i.e. relative search volume (RSV) were downloaded for this term using the all categories, web search and only YouTube settings during a period from 1.1.2019 to 3.31.2021 from www.trends.google.com. The geographic trends for this search query were plotted across the world.

Results

Seasonal peaks were observed for the search term “anatomy” during the first 3 months of the year and in months of September, October and November in 2019. Similar seasonal peaks were observed for the year 2020 except that there was sudden decrease in RSV for term “anatomy” in the month of March. Though trend for the rest of the year was same, but there was general lower RSV in 2020. The country with highest search hit was United States followed by Canada, Ireland, Australia and Philippines.

Conclusion

The search trend for pre-pandemic and pandemic period was similar with overall lower RSV during 2020, where it noticeably decreased during the initial phase of lockdown, i.e. in the month of March. As the whole world is still in the COVID-19 pandemic era, the future studies may report the google trends once the pandemic is over and may compare the post-pandemic trend for the same.

The predictive skill of convolutional neural networks models for disease forecasting

In this paper we investigate the utility of one-dimensional convolutional neural network (CNN) models in epidemiological forecasting. Deep learning models, in particular variants of recurrent neural networks (RNNs) have been studied for ILI (Influenza-Like Illness) forecasting, and have achieved a higher forecasting skill compared to conventional models such as ARIMA. In this study, we adapt two neural networks that employ one-dimensional temporal convolutional layers as a primary building block-temporal convolutional networks and simple neural attentive meta-learners-for epidemiological forecasting. We then test them with influenza data from the US collected over 2010-2019. We find that epidemiological forecasting with CNNs is feasible, and their forecasting skill is comparable to, and at times, superior to, plain RNNs. Thus CNNs and RNNs bring the power of nonlinear transformations to purely data-driven epidemiological models, a capability that heretofore has been limited to more elaborate mechanistic/compartmental disease models.

EagleEye: A Worldwide Disease-Related Topic Extraction System Using a Deep Learning Based Ranking Algorithm and Internet-Sourced Data

Due to the prevalence of globalization and the surge in people’s traffic, diseases are spreading more rapidly than ever and the risks of sporadic contamination are becoming higher than before. Disease warnings continue to rely on censored data, but these warning systems have failed to cope with the speed of disease proliferation. Due to the risks associated with the problem, there have been many studies on disease outbreak surveillance systems, but existing systems have limitations in monitoring disease-related topics and internationalization. With the advent of online news, social media and search engines, social and web data contain rich unexplored data that can be leveraged to provide accurate, timely disease activities and risks. In this study, we develop an infectious disease surveillance system for extracting information related to emerging diseases from a variety of Internet-sourced data. We also propose an effective deep learning-based data filtering and ranking algorithm. This system provides nation-specific disease outbreak information, disease-related topic ranking, a number of reports per district and disease through various visualization techniques such as a map, graph, chart, correlation and coefficient, and word cloud. Our system provides an automated web-based service, and it is free for all users and live in operation.

Accurate influenza forecasts using type-specific incidence data for small geographic units

Influenza incidence forecasting is used to facilitate better health system planning and could potentially be used to allow at-risk individuals to modify their behavior during a severe seasonal influenza epidemic or a novel respiratory pandemic. For example, the US Centers for Disease Control and Prevention (CDC) runs an annual competition to forecast influenza-like illness (ILI) at the regional and national levels in the US, based on a standard discretized incidence scale. Here, we use a suite of forecasting models to analyze type-specific incidence at the smaller spatial scale of clusters of nearby counties. We used data from point-of-care (POC) diagnostic machines over three seasons, in 10 clusters, capturing: 57 counties; 1,061,891 total specimens; and 173,909 specimens positive for Influenza A. Total specimens were closely correlated with comparable CDC ILI data. Mechanistic models were substantially more accurate when forecasting influenza A positive POC data than total specimen POC data, especially at longer lead times. Also, models that fit subpopulations of the cluster (individual counties) separately were better able to forecast clusters than were models that directly fit to aggregated cluster data. Public health authorities may wish to consider developing forecasting pipelines for type-specific POC data in addition to ILI data. Simple mechanistic models will likely improve forecast accuracy when applied at small spatial scales to pathogen-specific data before being scaled to larger geographical units and broader syndromic data. Highly local forecasts may enable new public health messaging to encourage at-risk individuals to temporarily reduce their social mixing during seasonal peaks and guide public health intervention policy during potentially severe novel influenza pandemics.

A novel data-driven methodology for influenza outbreak detection and prediction

Influenza is an infectious disease that leads to an estimated 5 million cases of severe illness and 650,000 respiratory deaths worldwide each year. The early detection and prediction of influenza outbreaks are crucial for efficient resource planning to save patient’s lives and healthcare costs. We propose a new data-driven methodology for influenza outbreak detection and prediction at very local levels. A doctor’s diagnostic dataset of influenza-like illness from more than 3000 clinics in Malaysia is used in this study because these diagnostic data are reliable and can be captured promptly. A new region index (RI) of the influenza outbreak is proposed based on the diagnostic dataset. By analysing the anomalies in the weekly RI value, potential outbreaks are identified using statistical methods. An ensemble learning method is developed to predict potential influenza outbreaks. Cross-validation is conducted to optimize the hyperparameters of the ensemble model. A testing data set is used to provide an unbiased evaluation of the model. The proposed methodology is shown to be sensitive and accurate at influenza outbreak prediction, with average of 75% recall, 74% precision, and 83% accuracy scores across five regions in Malaysia. The results are also validated by Google Flu Trends data, news reports, and surveillance data released by World Health Organization.

Predicting regional influenza epidemics with uncertainty estimation using commuting data in Japan

Obtaining an accurate prediction of the number of influenza patients in specific areas is a crucial task undertaken by medical institutions. Infections (such as influenza) spread from person to person, and people are rarely confined to a single area. Therefore, creating a regional influenza prediction model should consider the flow of people between different areas. Although various regional flu prediction models have previously been proposed, they do not consider the flow of people among areas. In this study, we propose a method that can predict the geographical distribution of influenza patients using commuting data to represent the flow of people. To elucidate the complex spatial dependence relations, our model uses an extension of the graph convolutional network (GCN). Additionally, a prediction interval for medical institutions is proposed, which is suitable for cyclic time series. Subsequently, we used the weekly data of flu patients from health authorities as the ground-truth to evaluate the prediction interval and performance of influenza patient prediction in each prefecture in Japan. The results indicate that our GCN-based model, which used commuting data, considerably improved the predictive accuracy over baseline values both temporally and spatially to provide an appropriate prediction interval. The proposed model is vital in practical settings, such as in the decision making of public health authorities and addressing growth in vaccine demand and workload. This paper primarily presents a GCN as a useful means for predicting the spread of an epidemic.

The development and deployment of a model for hospital-level COVID-19 associated patient demand intervals from consistent estimators (DICE)

Hospitals commonly project demand for their services by combining their historical share of regional demand with forecasts of total regional demand. Hospital-specific forecasts of demand that provide prediction intervals, rather than point estimates, may facilitate better managerial decisions, especially when demand overage and underage are associated with high, asymmetric costs. Regional point forecasts of patient demand are commonly available, e.g., for the number of people requiring hospitalization due to an epidemic such as COVID-19. However, even in this common setting, no probabilistic, consistent, computationally tractable forecast is available for the fraction of patients in a region that a particular institution should expect. We introduce such a forecast, DICE (Demand Intervals from Consistent Estimators). We describe its development and deployment at an academic medical center in California during the ‘second wave’ of COVID-19 in the Unite States. We show that DICE is consistent under mild assumptions and suitable for use with perfect, biased and unbiased regional forecasts. We evaluate its performance on empirical data from a large academic medical center as well as on synthetic data.

A computational tool for trend analysis and forecast of the COVID-19 pandemic

Purpose:

This paper proposes a methodology and a computational tool to study the COVID-19 pandemic throughout the world and to perform a trend analysis to assess its local dynamics.

Methods:

Mathematical functions are employed to describe the number of cases and demises in each region and to predict their final numbers, as well as the dates of maximum daily occurrences and the local stabilization date. The model parameters are calibrated using a computational methodology for numerical optimization. Trend analyses are run, allowing to assess the effects of public policies. Easy to interpret metrics over the quality of the fitted curves are provided. Country-wise data from the European Centre for Disease Prevention and Control (ECDC) concerning the daily number of cases and demises around the world are used, as well as detailed data from Johns Hopkins University and from the Brasil.io project describing individually the occurrences in United States counties and in Brazilian states and cities, respectively. U. S. and Brazil were chosen for a more detailed analysis because they are the current focus of the pandemic.

Results:

Illustrative results for different countries, U. S. counties and Brazilian states and cities are presented and discussed.

Conclusion:

The main contributions of this work lie in (i) a straightforward model of the curves to represent the data, which allows automation of the process without requiring interventions from experts; (ii) an innovative approach for trend analysis, whose results provide important information to support authorities in their decision-making process; and (iii) the developed computational tool, which is freely available and allows the user to quickly update the COVID-19 analyses and forecasts for any country, United States county or Brazilian state or city present in the periodic reports from the authorities.