VTechWorks Repository :: Browsing by Author "Lewis, Bryan L."

Browsing by Author "Lewis, Bryan L."

Now showing 1 - 20 of 28

Accuracy of epidemiological inferences based on publicly available information: retrospective comparative analysis of line lists of human cases infected with influenza A(H7N9) in China
Lau, Eric H. Y.; Zheng, Jiandong; Tsang, Tim K.; Liao, Qiaohong; Lewis, Bryan L.; Brownstein, John S.; Sanders, Sharon; Wong, Jessica Y.; Mekaru, Sumiko R.; Rivers, Caitlin; Wu, Peng; Jiang, Hui; Li, Yu; Yu, Jianxing; Zhang, Qian; Chang, Zhaorui; Liu, Fengfeng; Peng, Zhibin; Leung, Gabriel M.; Feng, Luzhao; Cowling, Benjamin J.; Yu, Hongjie (2014-05-28)
Background Appropriate public health responses to infectious disease threats should be based on best-available evidence, which requires timely reliable data for appropriate analysis. During the early stages of epidemics, analysis of ‘line lists’ with detailed information on laboratory-confirmed cases can provide important insights into the epidemiology of a specific disease. The objective of the present study was to investigate the extent to which reliable epidemiologic inferences could be made from publicly-available epidemiologic data of human infection with influenza A(H7N9) virus. Methods We collated and compared six different line lists of laboratory-confirmed human cases of influenza A(H7N9) virus infection in the 2013 outbreak in China, including the official line list constructed by the Chinese Center for Disease Control and Prevention plus five other line lists by HealthMap, Virginia Tech, Bloomberg News, the University of Hong Kong and FluTrackers, based on publicly-available information. We characterized clinical severity and transmissibility of the outbreak, using line lists available at specific dates to estimate epidemiologic parameters, to replicate real-time inferences on the hospitalization fatality risk, and the impact of live poultry market closure. Results Demographic information was mostly complete (less than 10% missing for all variables) in different line lists, but there were more missing data on dates of hospitalization, discharge and health status (more than 10% missing for each variable). The estimated onset to hospitalization distributions were similar (median ranged from 4.6 to 5.6 days) for all line lists. Hospital fatality risk was consistently around 20% in the early phase of the epidemic for all line lists and approached the final estimate of 35% afterwards for the official line list only. Most of the line lists estimated >90% reduction in incidence rates after live poultry market closures in Shanghai, Nanjing and Hangzhou. Conclusions We demonstrated that analysis of publicly-available data on H7N9 permitted reliable assessment of transmissibility and geographical dispersion, while assessment of clinical severity was less straightforward. Our results highlight the potential value in constructing a minimum dataset with standardized format and definition, and regular updates of patient status. Such an approach could be particularly useful for diseases that spread across multiple countries.
Advancing the Global Land Grant Institution: Creating a Virtual Environment to Re-envision Extension and Advance GSS-related Research, Education, and Collaboration
Hall, Ralph P.; Polys, Nicholas F.; Sforza, Peter M.; Eubank, Stephen D.; Lewis, Bryan L.; Krometis, Leigh-Anne H.; Pollyea, Ryan M.; Schoenholtz, Stephen H.; Sridhar, Venkataramana; Crowder, Van; Lipsey, John; Christie, Maria Elisa; Glasson, George E.; Scherer, Hannah H.; Davis, A. Jack; Dunay, Robert J.; King, Nathan T.; Muelenaer, Andre A.; Muelenaer, Penelope; Rist, Cassidy; Wenzel, Sophie (Virginia Tech, 2017-05-15)
The vision for this project has emerged from several years of research, teaching, and service in Africa and holds the potential to internationalize education at Virginia Tech and in our partner institutions in Malawi. The vision is simple, to develop a state-of-the-art, data rich, virtual decision-support and learning environment that enables local-, regional-, and national-level actors in developed and developing regions to make decisions that improve resilience and sustainability. Achieving these objectives will require a system that can combine biogeophysical and sociocultural data in a way that enables actors to understand and leverage these data to enhance decision-making at various levels. The project will begin by focusing on water, agricultural, and health systems in Malawi, and can be expanded over time to include any sector or system in any country. The core ideas are inherently scalable...
Applying Time-Valued Knowledge for Public Health Outbreak Response
Schlitt, James Thomas (Virginia Tech, 2019-06-21)
During the early stages of any epidemic, simple interventions such as quarantine and isolation may be sufficient to halt the spread of a novel pathogen. However, should this opportunity be missed, substantially more resource-intensive, complex, and societally intrusive interventions may be required to achieve an acceptable outcome. These disparities place a differential on the value of a given unit of knowledge across the time-domains of an epidemic. Within this dissertation we explore these value-differentials via extension of the business concept of the time-value of knowledge and propose the C4 Response Model for organizing the research response to novel pathogenic outbreaks. First, we define the C4 Response Model as a progression from an initial data-hungry collect stage, iteration between open-science-centric connect stages and machine-learning centric calibrate stages, and a final visualization-centric convey stage. Secondly we analyze the trends in knowledge-building across the stages of epidemics with regard to open and closed access article publication, referencing, and citation. Thirdly, we demonstrate a Twitter message mapping application to assess the virality of tweets as a function of their source-profile category, message category, timing, urban context, tone, and use of bots. Finally, we apply an agent-based model of influenza transmission to explore the efficacy of combined antiviral, sequestration, and vaccination interventions in mitigating an outbreak of an influenza-like-illness (ILI) within a simulated military base population. We find that while closed access outbreak response articles use more recent citations and see higher mean citation counts, open access articles are published and referenced in significantly greater numbers and are growing in proportion. We observe that tweet viralities showed distinct heterogeneities across message and profile type pairing, that tweets dissipated rapidly across time and space, and that tweets published before high-tweet-volume time periods showed higher virality. Finally, we saw that while timely responses and strong pharmaceutical interventions showed the greatest impact in mitigating ILI transmission within a military base, even optimistic scenarios failed to prevent the majority of new cases. This body of work offers significant methodological contributions for the practice of computational epidemiology as well as a theoretical grounding for the further use of the C4 Response Model.
Comparing Effectiveness of Top-Down and Bottom-Up Strategies in Containing Influenza
Marathe, Achla; Lewis, Bryan L.; Barrett, Christopher L.; Chen, Jiangzhuo; Marathe, Madhav V.; Eubank, Stephen; Ma, Yifei (Public Library of Science, 2011-09-22)
This research compares the performance of bottom-up, self-motivated behavioral interventions with top-down interventions targeted at controlling an “Influenza-like-illness”. Both types of interventions use a variant of the ring strategy. In the first case, when the fraction of a person's direct contacts who are diagnosed exceeds a threshold, that person decides to seek prophylaxis, e.g. vaccine or antivirals; in the second case, we consider two intervention protocols, denoted Block and School: when a fraction of people who are diagnosed in a Census Block (resp., School) exceeds the threshold, prophylax the entire Block (resp., School). Results show that the bottom-up strategy outperforms the top-down strategies under our parameter settings. Even in situations where the Block strategy reduces the overall attack rate well, it incurs a much higher cost. These findings lend credence to the notion that if people used antivirals effectively, making them available quickly on demand to private citizens could be a very effective way to control an outbreak.
Detail in network models of epidemiology: are we there yet?
Eubank, Stephen; Barrett, Christopher L.; Beckman, Richard J.; Bisset, Keith R.; Durbeck, L.; Kuhlman, Christopher J.; Lewis, Bryan L.; Marathe, Achla; Marathe, Madhav V.; Stretz, P. (Taylor & Francis, 2010)
Network models of infectious disease epidemiology can potentially provide insight into how to tailor control strategies for specific regions, but only if the network adequately reflects the structure of the region’s contact network. Typically, the network is produced by models that incorporate details about human interactions. Each detail added renders the models more complicated and more difficult to calibrate, but also more faithful to the actual contact network structure. We propose a statistical test to determine when sufficient detail has been added to the models and demonstrate its application to the models used to create a synthetic population and contact network for the USA.
Disparities in spread and control of influenza in slums of Delhi: findings from an agent-based modelling study
Adiga, Abhijin; Chu, Shuyu; Kuhlman, Christopher J.; Lewis, Bryan L.; Marathe, Achla; Nordberg, Eric K.; Swarup, Samarth; Vullikanti, Anil; Wilson, Mandy L. (BMJ Publishing Group, 2017-11-03)
Objectives: This research studies the role of slums in the spread and control of infectious diseases in the National Capital Territory of India, Delhi, using detailed social contact networks of its residents. Methods: We use an agent-based model to study the spread of influenza in Delhi through person-to-person contact. Two different networks are used: one in which slum and non-slum regions are treated the same, and the other in which 298 slum zones are identified. In the second network, slum-specific demographics and activities are assigned to the individuals whose homes reside inside these zones. The main effects of integrating slums are that the network has more home-related contacts due to larger family sizes and more outside contacts due to more daily activities outside home. Various vaccination and social distancing interventions are applied to control the spread of influenza. Results: Simulation-based results show that when slum attributes are ignored, the effectiveness of vaccination can be overestimated by 30%–55%, in terms of reducing the peak number of infections and the size of the epidemic, and in delaying the time to peak infection. The slum population sustains greater infection rates under all intervention scenarios in the network that treats slums differently. Vaccination strategy performs better than social distancing strategies in slums. Conclusions: Unique characteristics of slums play a significant role in the spread of infectious diseases. Modelling slums and estimating their impact on epidemics will help policy makers and regulators more accurately prioritise allocation of scarce medical resources and implement public health policies.
Epidemiological and economic impact of pandemic influenza in Chicago: Priorities for vaccine interventions
Dorratoltaj, Nargesalsadat; Marathe, Achla; Lewis, Bryan L.; Swarup, Samarth; Eubank, Stephen G.; Abbas, Kaja M. (PLOS, 2017-06-01)
The study objective is to estimate the epidemiological and economic impact of vaccine interventions during influenza pandemics in Chicago, and assist in vaccine intervention priorities. Scenarios of delay in vaccine introduction with limited vaccine efficacy and limited supplies are not unlikely in future influenza pandemics, as in the 2009 H1N1 influenza pandemic. We simulated influenza pandemics in Chicago using agent-based transmission dynamic modeling. Population was distributed among high-risk and non-high risk among 0±19, 20±64 and 65+ years subpopulations. Different attack rate scenarios for catastrophic (30.15%), strong (21.96%), and moderate (11.73%) influenza pandemics were compared against vaccine intervention scenarios, at 40% coverage, 40% efficacy, and unit cost of $28.62. Sensitivity analysis for vaccine compliance, vaccine efficacy and vaccine start date was also conducted. Vaccine prioritization criteria include risk of death, total deaths, net benefits, and return on investment. The risk of death is the highest among the high-risk 65+ years subpopulation in the catastrophic influenza pandemic, and highest among the high-risk 0±19 years subpopulation in the strong and moderate influenza pandemics. The proportion of total deaths and net benefits are the highest among the high-risk 20±64 years subpopulation in the catastrophic, strong and moderate influenza pandemics. The return on investment is the highest in the high-risk 0±19 years subpopulation in the catastrophic, strong and moderate influenza pandemics. Based on risk of death and return on investment, high-risk groups of the three age group subpopulations can be prioritized for vaccination, and the vaccine interventions are cost saving for all age and risk groups. The attack rates among the children are higher than among the adults and seniors in the catastrophic, strong, and moderate influenza pandemic scenarios, due to their larger social contact network and homophilous interactions in school. Based on return on investment and higher attack rates among children, we recommend prioritizing children (0±19 years) and seniors (65+ years) after high-risk groups for influenza vaccination during times of limited vaccine supplies. Based on risk of death, we recommend prioritizing seniors (65+ years) after high-risk groups for influenza vaccination during times of limited vaccine supplies.
EpiViewer: an epidemiological application for exploring time series data
Thorve, Swapna; Wilson, Mandy L.; Lewis, Bryan L.; Swarup, Samarth; Vullikanti, Anil Kumar S.; Marathe, Madhav V. (2018-11-22)
Background Visualization plays an important role in epidemic time series analysis and forecasting. Viewing time series data plotted on a graph can help researchers identify anomalies and unexpected trends that could be overlooked if the data were reviewed in tabular form; these details can influence a researcher’s recommended course of action or choice of simulation models. However, there are challenges in reviewing data sets from multiple data sources – data can be aggregated in different ways (e.g., incidence vs. cumulative), measure different criteria (e.g., infection counts, hospitalizations, and deaths), or represent different geographical scales (e.g., nation, HHS Regions, or states), which can make a direct comparison between time series difficult. In the face of an emerging epidemic, the ability to visualize time series from various sources and organizations and to reconcile these datasets based on different criteria could be key in developing accurate forecasts and identifying effective interventions. Many tools have been developed for visualizing temporal data; however, none yet supports all the functionality needed for easy collaborative visualization and analysis of epidemic data. Results In this paper, we present EpiViewer, a time series exploration dashboard where users can upload epidemiological time series data from a variety of sources and compare, organize, and track how data evolves as an epidemic progresses. EpiViewer provides an easy-to-use web interface for visualizing temporal datasets either as line charts or bar charts. The application provides enhanced features for visual analysis, such as hierarchical categorization, zooming, and filtering, to enable detailed inspection and comparison of multiple time series on a single canvas. Finally, EpiViewer provides several built-in statistical Epi-features to help users interpret the epidemiological curves. Conclusions EpiViewer is a single page web application that provides a framework for exploring, comparing, and organizing temporal datasets. It offers a variety of features for convenient filtering and analysis of epicurves based on meta-attribute tagging. EpiViewer also provides a platform for sharing data between groups for better comparison and analysis. Our user study demonstrated that EpiViewer is easy to use and fills a particular niche in the toolspace for visualization and exploration of epidemiological data.
Estimating Human Cases of Avian Influenza A(H7N9) from Poultry Exposure
Rivers, Caitlin; Lum, Kristian; Lewis, Bryan L.; Eubank, Stephen (PLOS, 2013-05-15)
In March 2013 an outbreak of avian influenza A(H7N9) was first recognized in China. To date there have been 130 cases in human, 47% of which are in men over the age of 55.The influenza strain is a novel subtype not seen before in humans; little is known about zoonotic transmission of the virus, but it is hypothesized that contact with poultry in live bird markets may be a source of exposure. The purpose of this study is to estimate the transmissibility of the virus from poultry to humans by estimating the amount of time shoppers, farmers, and live bird market retailers spend exposed to poultry each day. Results suggest that increased risk among older men is not due to greater exposure time at live bird markets.
Forecasting influenza activity using machine-learned mobility map
Venkatramanan, Srinivasan; Sadilek, Adam; Fadikar, Arindam; Barrett, Christopher L.; Biggerstaff, Matthew; Chen, Jiangzhuo; Dotiwalla, Xerxes; Eastham, Paul; Gipson, Bryant; Higdon, Dave; Kucuktunc, Onur; Lieber, Allison; Lewis, Bryan L.; Reynolds, Zane; Vullikanti, Anil Kumar S.; Wang, Lijing; Marathe, Madhav V. (2021-02-09)
Human mobility is a primary driver of infectious disease spread. However, existing data is limited in availability, coverage, granularity, and timeliness. Data-driven forecasts of disease dynamics are crucial for decision-making by health officials and private citizens alike. In this work, we focus on a machine-learned anonymized mobility map (hereon referred to as AMM) aggregated over hundreds of millions of smartphones and evaluate its utility in forecasting epidemics. We factor AMM into a metapopulation model to retrospectively forecast influenza in the USA and Australia. We show that the AMM model performs on-par with those based on commuter surveys, which are sparsely available and expensive. We also compare it with gravity and radiation based models of mobility, and find that the radiation model's performance is quite similar to AMM and commuter flows. Additionally, we demonstrate our model's ability to predict disease spread even across state boundaries. Our work contributes towards developing timely infectious disease forecasting at a global scale using human mobility datasets expanding their applications in the area of infectious disease epidemiology. Human mobility plays a central role in the spread of infectious diseases and can help in forecasting incidence. Here the authors show a comparison of multiple mobility benchmarks in forecasting influenza, and demonstrate the value of a machine-learned mobility map with global coverage at multiple spatial scales.
A framework for evaluating epidemic forecasts
Tabataba, Farzaneh Sadat; Chakraborty, Prithwish; Ramakrishnan, Naren; Venkatramanan, Srinivasan; Chen, Jiangzhuo; Lewis, Bryan L.; Marathe, Madhav V. (2017-05-15)
Background Over the past few decades, numerous forecasting methods have been proposed in the field of epidemic forecasting. Such methods can be classified into different categories such as deterministic vs. probabilistic, comparative methods vs. generative methods, and so on. In some of the more popular comparative methods, researchers compare observed epidemiological data from the early stages of an outbreak with the output of proposed models to forecast the future trend and prevalence of the pandemic. A significant problem in this area is the lack of standard well-defined evaluation measures to select the best algorithm among different ones, as well as for selecting the best possible configuration for a particular algorithm. Results In this paper we present an evaluation framework which allows for combining different features, error measures, and ranking schema to evaluate forecasts. We describe the various epidemic features (Epi-features) included to characterize the output of forecasting methods and provide suitable error measures that could be used to evaluate the accuracy of the methods with respect to these Epi-features. We focus on long-term predictions rather than short-term forecasting and demonstrate the utility of the framework by evaluating six forecasting methods for predicting influenza in the United States. Our results demonstrate that different error measures lead to different rankings even for a single Epi-feature. Further, our experimental analyses show that no single method dominates the rest in predicting all Epi-features when evaluated across error measures. As an alternative, we provide various Consensus Ranking schema that summarize individual rankings, thus accounting for different error measures. Since each Epi-feature presents a different aspect of the epidemic, multiple methods need to be combined to provide a comprehensive forecast. Thus we call for a more nuanced approach while evaluating epidemic forecasts and we believe that a comprehensive evaluation framework, as presented in this paper, will add value to the computational epidemiology community.
in silico Public Health: The Essential Role of Highly Detailed Simulations in Support of Public Health Decision-Making
Lewis, Bryan L. (Virginia Tech, 2011-01-19)
Public Health requires a trans-disciplinary approach to tackle the breadth and depth of the issues it faces. Public health decisions are reached through the compilation of multiple data sources and their thoughtful synthesis. The complexity and importance of these decisions necessitates a variety of approaches, with simulations increasingly being relied upon. This dissertation describes several research efforts that demonstrate the utility of highly detailed simulations in public health decision-making. Simulations are frequently used to represent dynamic processes and to synthesize data to predict future outcomes, which can be used in cost-benefit and course of action analyses. The threat of pandemic influenza and its subsequent arrival prompted many simulation-based studies. This dissertation details several such studies conducted at the federal policy level. Their use for planning and the rapid response to the unfolding crisis demonstrates the integration of highly detailed simulations into the public health decision-making process. Most analytic methods developed by public health practitioners rely on historical data sources, but are intended to be broadly applicable. Oftentimes this data is limited or incomplete. This dissertation describes the use of highly detailed simulations to evaluate the performance of outbreak detection algorithms. By creating methods that generate realistic and configurable synthetic data, the reliance on these historical samples can be reduced, thus facilitating the development and improvement of methods for public health practice. The process of decision-making itself can significantly influence the decisions reached. Many fields use simulations to train and evaluate, however, public health has yet to fully adopt these approaches. This dissertation details the construction of highly detailed synthetic data that was used to build an interactive environment designed to evaluate the decision-making processes for pertussis control. The realistic data sets provide sufficient face validity to experienced public health practitioners, creating a natural and effective medium for training and evaluation purposes. Advances in high-performance computing, information sciences, computer science, and epidemiology are enabling increasing innovation in the application of simulations. This dissertation illustrates several applications of simulations to relevant public health practices and strongly argues that highly detailed simulations have an essential role to play in Public Health decision-making.
Lyme Disease and Forest Fragmentation in the Peridomestic Environment
Telionis, Pyrros A. (Virginia Tech, 2020-05-14)
Over the last 20 years, Lyme disease has grown to become the most common vector-borne disease affecting Americans. Spread in the eastern U.S. primarily by the bite of Ixodes scapularis, the black-legged tick, the disease affects an estimated 329,000 Americans per year. Originally confined to New England, it has since spread across much of the east coast and has become endemic in Virginia. Since 2010 the state has averaged 1200 cases per year, with 200 annually in the New River Health District (NRHD), the location of our study. Efforts to geographically model Lyme disease primarily focus on landscape and climatic variables. The disease depends highly on the survival of the tick vector, and white-footed mouse, the primary reservoir. Both depend on the existence of forest-herbaceous edge-habitats, as well as warm summer temperatures, mild winter lows, and summer wetness. While many studies have investigated the effect of forest fragmentation on Lyme, none have made use of high-resolution land cover data to do so at the peridomestic level. To fill this knowledge gap, we made use of the Virginia Geographic Information Network’s 1-meter land cover dataset and identified forest-herbaceous edge-habitats for the NRHD. We then calculated the density of these edge-habitats at 100, 200 and 300-meter radii, representing the peridomestic environment. We also calculated the density of <2-hectare forest patches at the same distance thresholds. To avoid confounding from climatic variation, we also calculated mean summer temperatures, total summer rainfall, and number of consecutive days below freezing of the prior winters. Adding to these data, elevation, terrain shape index, slope, and aspect, and including lags on each of our climatic variables, we created environmental niche models of Lyme in the NRHD. We did so using both Boosted Regression Trees (BRT) and Maximum Entropy (MaxEnt) modeling, the two most common niche modeling algorithms in the field today. We found that Lyme is strongly associated with higher density of developed-herbaceous edges within 100-meters from the home. Forest patch density was also significant at both 100-meter and 300-meter levels. This supports the notion that the fine scale peridomestic environment is significant to Lyme outcomes, and must be considered even if one were to account for fragmentation at a wider scale, as well as variations in climate and terrain.
Modeling Emerging Infectious Diseases for Public Health Decision Support
Rivers, Caitlin (Virginia Tech, 2015-05-05)
Emerging infectious diseases (EID) pose a serious threat to global public health. Computational epidemiology is a nascent subfield of public health that can provide insight into an outbreak in advance of traditional methodologies. Research in this dissertation will use fuse nontraditional, publicly available data sources with more traditional epidemiological data to build and parameterize models of emerging infectious diseases. These methods will be applied to avian influenza A (H7N9), Middle Eastern Respiratory Syndrome Coronavirus (MERS-CoV), and Ebola virus disease (EVD) outbreaks. This effort will provide quantitative, evidenced-based guidance for policymakers and public health responders to augment public health operations.
Modeling targeted layered containment of an influenza pandemic in the United States
Halloran, Elizabeth M.; Ferguson, Neil M.; Eubank, Stephen; Longini, Ira M. Jr.; Cummings, Derek A. T.; Lewis, Bryan L.; Xu, Shufu; Fraser, Christophe; Vullikanti, Anil; Germann, Timothy C.; Wagener, Diane; Beckman, Richard J.; Kadau, Kai; Barrett, Christopher L.; Macken, Catherine A.; Burke, Donald S.; Cooley, Philip (NAS, 2008-03-25)
Planning a response to an outbreak of a pandemic strain of influenza is a high public health priority. Three research groups using different individual-based, stochastic simulation models have examined the consequences of intervention strategies chosen in consultation with U.S. public health workers. The first goal is to simulate the effectiveness of a set of potentially feasible intervention strategies. Combinations called targeted layered containment (TLC) of influenza antiviral treatment and prophylaxis and nonpharmaceutical interventions of quarantine, isolation, school closure, community social distancing, and workplace social distancing are considered. The second goal is to examine the robustness of the results to model assumptions. The comparisons focus on a pandemic outbreak in a population similar to that of Chicago, with 8.6 million people. The simulations suggest that at the expected transmissibility of a pandemic strain, timely implementation of a combination of targeted household antiviral prophylaxis, and social distancing measures could substantially lower the illness attack rate before a highly efficacious vaccine could become available. Timely initiation of measures and school closure play important roles. Because of the current lack of data on which to base such models, further field research is recommended to learn more about the sources of transmission and the effectiveness of social distancing measures in reducing influenza transmission.
Modeling the Ebola Outbreak in West Africa, August 4th 2014 update
Lewis, Bryan L.; Rivers, Caitlin; Eubank, Stephen; Marathe, Marathe; Barrett, Christopher L. (2014)
News Analytics for Global Infectious Disease Surveillance
Ghosh, Saurav (Virginia Tech, 2017-11-29)
Traditional disease surveillance can be augmented with a wide variety of open sources, such as online news media, twitter, blogs, and web search records. Rapidly increasing volumes of these open sources are proving to be extremely valuable resources in helping analyze, detect, and forecast outbreaks of infectious diseases, especially new diseases or diseases spreading to new regions. However, these sources are in general unstructured (noisy) and construction of surveillance tools ranging from real-time disease outbreak monitoring to construction of epidemiological line lists involves considerable human supervision. Intelligent modeling of such sources using text mining methods such as, topic models, deep learning and dependency parsing can lead to automated generation of the mentioned surveillance tools. Moreover, real-time global availability of these open sources from web-based bio-surveillance systems, such as HealthMap and WHO Disease Outbreak News (DONs) can aid in development of generic tools which will be applicable to a wide range of diseases (rare, endemic and emerging) across different regions of the world. In this dissertation, we explore various methods of using internet news reports to develop generic surveillance tools which can supplement traditional surveillance systems and aid in early detection of outbreaks. We primarily investigate three major problems related to infectious disease surveillance as follows. (i) Can trends in online news reporting monitor and possibly estimate infectious disease outbreaks? We introduce approaches that use temporal topic models over HealthMap corpus for detecting rare and endemic disease topics as well as capturing temporal trends (seasonality, abrupt peaks) for each disease topic. The discovery of temporal topic trends is followed by time-series regression techniques to estimate future disease incidence. (ii) In the second problem, we seek to automate the creation of epidemiological line lists for emerging diseases from WHO DONs in a near real-time setting. For this purpose, we formulate Guided Epidemiological Line List (GELL), an approach that combines neural word embeddings with information extracted from dependency parse-trees at the sentence level to extract line list features. (iii) Finally, for the third problem, we aim to characterize diseases automatically from HealthMap corpus using a disease-specific word embedding model which were subsequently evaluated against human curated ones for accuracies.
Novel Applications of Geospatial Analysis in the Modeling of Infectious Diseases
Telionis, Pyrros A. (Virginia Tech, 2019-05-08)
At the intersection of geography and public health, the field of spatial epidemiology seeks to use the tools of geospatial analysis to answer questions about disease. In this work we explore two areas: the use of geostatistical modeling as an extension of niche modeling, and the use of mobility metrics to augment modeling for epidemic responses. Niche modeling refers to the practice of using statistical methods to relate the underlying spatially distributed environmental variables to an outcome, typically presence or absence of a species. Such work is common in disease ecology, and often focuses on exploring the range of a disease vector or pathogen. The technique also allows one to explore the importance of each underlying regressor, and the effect it has on the outcome. We demonstrate that this concept can be extended, through geostatistical modeling, to explore non-logistic phenomena such as incidence. When combined with weather forecasts, such efforts can even predict incidence of an upcoming season, allowing us to estimate the total number of expected cases, and where we would expect to find them. We demonstrate this in Chapter 2, by forecasting the incidence of melioidosis in Australia given weather forecasts a year prior. We also evaluate the efficacy of this technique and explore the impact of environmental variables such as elevation on melioidosis. But these techniques are not limited to free-living and vector-borne pathogens. We theorize that they can also be applied to diseases that spread exclusively by person-to-person contact. Exploring this allows us to find areas of underreporting, as well as areas with unusual local forcing which might merit further investigation by the health department. We also explore this in Chapter 4, by relating the incidence of hepatitis C in rural Virginia to demographic data. The West African Ebola Outbreak of 2014 demonstrated the need to include mobility in predictive disease modeling. One can no longer assume that neglected tropical diseases will remain contained and immobile, and the assumption of random mixing across large areas is unwise. Our efforts with modeling mobility are twofold. In Chapter 3, we demonstrate the creation of mobility metrics from open source road and river network data. We then demonstrate the usefulness of such data in a meta-population patch model meant to forecast the spread of Ebola in the Democratic Republic of Congo. In Chapter 4, we also demonstrate that mobility data can be used to strengthen outbreak detection via hotspot analysis, and to augment incidence models by factoring in the incidence rates of neighboring areas. These efforts will allow health departments to more accurately forecast incidence, and more readily identify disease hotspots of atypical size and shape.
On the 3 M's of Epidemic Forecasting: Methods, Measures, and Metrics
Tabataba, Farzaneh Sadat (Virginia Tech, 2017-12-06)
Over the past few decades, various computational and mathematical methodologies have been proposed for forecasting seasonal epidemics. In recent years, the deadly effects of enormous pandemics such as the H1N1 influenza virus, Ebola, and Zika, have compelled scientists to find new ways to improve the reliability and accuracy of epidemic forecasts. The improvement and variety of these prediction methods are undeniable. Nevertheless, many challenges remain unresolved in the path of forecasting the outbreaks using surveillance data. Obtaining the clean real-time data has always been an obstacle. Moreover, the surveillance data is usually noisy and handling the uncertainty of the observed data is a major issue for forecasting algorithms. Correct modeling assumptions regarding the nature of the infectious disease is another dilemma. Oversimplified models could lead to inaccurate forecasts, whereas more complicated methods require additional computational resources and information. Without those, the model may not be able to converge to a unique optimum solution. Through the last decade, there has been a significant effort towards achieving better epidemic forecasting algorithms. However, the lack of standard, well-defined evaluating metrics impedes a fair judgment on the proposed methods. This dissertation is divided into two parts. In the first part, we present a Bayesian particle filter calibration framework integrated with an agent-based model to forecast the epidemic trend of diseases like flu and Ebola. Our approach uses Bayesian statistics to estimate the underlying disease model parameters given the observed data and handle the uncertainty in the reasoning. An individual-based model with different intervention strategies could result in a large number of unknown parameters that should be properly calibrated. As particle filter could collapse in very large-scale systems (curse-of-dimensionality problem), achieving the optimum solution becomes more challenging. Our proposed particle filter framework utilizes machine learning concepts to restrain the intractable search space. It incorporates a smart analyzer in the state dynamics unit that examines the predicted and observed data using machine learning techniques to guide the direction and amount of perturbation of each parameter in the searching process. The second part of this dissertation focuses on providing standard evaluation measures for evaluating epidemic forecasts. We present an end-to-end framework that introduces epidemiologically relevant features (Epi-features), error measures, and ranking schema as the main modules of the evaluation process. Lastly, we provide the evaluation framework as a software package named Epi-Evaluator and demonstrate the potentials and capabilities of the framework by applying it to the output of different forecasting methods.
Opinion: Mathematical models: A key tool for outbreak response
Lofgren, Eric T.; Halloran, M. Elizabeth; Rivers, Caitlin; Drake, John M.; Porco, Travis C.; Lewis, Bryan L.; Yang, Wan; Vespignani, Alessandro; Shaman, Jeffrey; Eisenberg, Joseph N.S.; Eisenberg, Marisa C.; Marathe, Madhav V.; Scarpino, Samuel V.; Alexander, Kathleen A.; Meza, Rafael; Ferrari, Matthew J.; Hyman, James M.; Meyers, Lauren Ancel; Eubank, Stephen (NAS, 2015-01-13)
The 2014 outbreak of Ebola in West Africa is unprecedented in its size and geographic range, and demands swift, effective action from the international community. Understanding the dynamics and spread of Ebola is critical for directing interventions and extinguishing the epidemic; however, observational studies of local conditions have been incomplete and limited by the urgent need to direct resources to patient care. Mathematical and computational models can help address this deficiency through work with sparse observations, inference on missing data, and incorporation of the latest information. These models can clarify how the disease is spreading and provide timely guidance to policymakers. However, the use of models in public health often meets resistance (1), from doubts in peer review about the utility of such analyses to public skepticism that models can contribute when the means to control an epidemic are already known (2). Even when they are discussed in a positive light, models are often portrayed as arcane and largely inaccessible thought experiments (3). However, the role of models is crucial: they can be used to quantify the effect of mitigation efforts, provide guidance on the scale of interventions required to achieve containment, and identify factors that fundamentally influence the course of an outbreak.

Browsing by Author "Lewis, Bryan L."

Results Per Page

Sort Options