Six challenges in measuring contact networks for use in modelling

Leveraging advances in data-driven deep learning methods for hybrid epidemic modeling

Mathematical modeling of epidemic dynamics is crucial to understand its underlying mechanisms, quantify important parameters, and make predictions that facilitate more informed decision-making. There are three major types of models: mechanistic models including the SEIR-type paradigm, alternative data-driven (DD) approaches, and hybrid models that combine mechanistic models with DD approaches. In this paper, we summarize our work in the COVID-19 Scenario Modeling Hub (SMH) for more than 12 rounds since early 2021 for informed decision support. We emphasize the importance of deep learning techniques for epidemic modeling via a flexible DD framework that substantially complements the mechanistic paradigm to evaluate various future epidemic scenarios. We start with a traditional curve-fitting approach to model cumulative COVID-19 based on the underlying SEIR-type mechanisms. Hospitalizations and deaths are modeled as binomial processes of cases and hospitalization, respectively. We further formulate two types of deep learning models based on multivariate long short term memory (LSTM) to address the challenges of more traditional DD models. The first LSTM is structurally similar to the curve fitting approach and assumes that hospitalizations and deaths are binomial processes of cases. Instead of using a predefined exponential curve, LSTM relies on the underlying data to identify the most appropriate functions, and is capable of capturing both long-term and short-term epidemic behaviors. We then relax the assumption of dependent inputs among cases, hospitalizations, and death. Another type of LSTM that handles all input time series as parallel signals, the independent multivariate LSTM, is developed. Independent multivariate LSTM can incorporate a wide range of data sources beyond traditional case-based epidemiological surveillance. The DD framework unleashes its potential in big data era with previously neglected heterogeneous surveillance data sources, such as syndromic, environment, genomic, serologic, infoveillance, and mobility data. DD approaches, especially LSTM, complement and integrate with the mechanistic modeling paradigm, provide a feasible alternative approach to model today's complex socio-epidemiological systems, and further leverage our ability to explore different scenarios for more informed decision-making during health emergencies.

An algorithm to build synthetic temporal contact networks based on close-proximity interactions data

Small populations (e.g., hospitals, schools or workplaces) are characterised by high contact heterogeneity and stochasticity affecting pathogen transmission dynamics. Empirical individual contact data provide unprecedented information to characterize such heterogeneity and are increasingly available, but are usually collected over a limited period, and can suffer from observation bias. We propose an algorithm to stochastically reconstruct realistic temporal networks from individual contact data in healthcare settings (HCS) and test this approach using real data previously collected in a long-term care facility (LTCF). Our algorithm generates full networks from recorded close-proximity interactions, using hourly inter-individual contact rates and information on individuals' wards, the categories of staff involved in contacts, and the frequency of recurring contacts. It also provides data augmentation by reconstructing contacts for days when some individuals are present in the HCS without having contacts recorded in the empirical data. Recording bias is formalized through an observation model, to allow direct comparison between the augmented and observed networks. We validate our algorithm using data collected during the i-Bird study, and compare the empirical and reconstructed networks. The algorithm was substantially more accurate to reproduce network characteristics than random graphs. The reconstructed networks reproduced well the assortativity by ward (first-third quartiles observed: 0.54-0.64; synthetic: 0.52-0.64) and the hourly staff and patient contact patterns. Importantly, the observed temporal correlation was also well reproduced (0.39-0.50 vs 0.37-0.44), indicating that our algorithm could recreate a realistic temporal structure. The algorithm consistently recreated unobserved contacts to generate full reconstructed networks for the LTCF. To conclude, we propose an approach to generate realistic temporal contact networks and reconstruct unobserved contacts from summary statistics computed using individual-level interaction networks. This could be applied and extended to generate contact networks to other HCS using limited empirical data, to subsequently inform individual-based epidemic models.

The effectiveness of intervention measures on MERS-CoV transmission by using the contact networks reconstructed from link prediction data

Introduction

Mitigating the spread of infectious diseases is of paramount concern for societal safety, necessitating the development of effective intervention measures. Epidemic simulation is widely used to evaluate the efficacy of such measures, but realistic simulation environments are crucial for meaningful insights. Despite the common use of contact-tracing data to construct realistic networks, they have inherent limitations. This study explores reconstructing simulation networks using link prediction methods as an alternative approach.

Methods

The primary objective of this study is to assess the effectiveness of intervention measures on the reconstructed network, focusing on the 2015 MERS-CoV outbreak in South Korea. Contact-tracing data were acquired, and simulation networks were reconstructed using the graph autoencoder (GAE)-based link prediction method. A scale-free (SF) network was employed for comparison purposes. Epidemic simulations were conducted to evaluate three intervention strategies: Mass Quarantine (MQ), Isolation, and Isolation combined with Acquaintance Quarantine (AQ + Isolation).

Results

Simulation results showed that AQ + Isolation was the most effective intervention on the GAE network, resulting in consistent epidemic curves due to high clustering coefficients. Conversely, MQ and AQ + Isolation were highly effective on the SF network, attributed to its low clustering coefficient and intervention sensitivity. Isolation alone exhibited reduced effectiveness. These findings emphasize the significant impact of network structure on intervention outcomes and suggest a potential overestimation of effectiveness in SF networks. Additionally, they highlight the complementary use of link prediction methods.

Discussion

This innovative methodology provides inspiration for enhancing simulation environments in future endeavors. It also offers valuable insights for informing public health decision-making processes, emphasizing the importance of realistic simulation environments and the potential of link prediction methods.

The spectral underpinnings of pathogen spread on animal networks

Predicting what factors promote or protect populations from infectious disease is a fundamental epidemiological challenge. Social networks, where nodes represent hosts and edges represent direct or indirect contacts between them, are important in quantifying these aspects of infectious disease dynamics. However, how network structure and epidemic parameters interact in empirical networks to promote or protect animal populations from infectious disease remains a challenge. Here we draw on advances in spectral graph theory and machine learning to build predictive models of pathogen spread on a large collection of empirical networks from across the animal kingdom. We show that the spectral features of an animal network are powerful predictors of pathogen spread for a variety of hosts and pathogens and can be a valuable proxy for the vulnerability of animal networks to pathogen spread. We validate our findings using interpretable machine learning techniques and provide a flexible web application for animal health practitioners to assess the vulnerability of a particular network to pathogen spread.

Network canvas: an open-source tool for capturing social and contact network data

MOTIVATION

Social influence and contact networks are extremely important for understanding health behaviour and the spread of disease. Yet, most traditional software tools are not optimized to capture these data, making measurement of personal networks challenging. Our team developed Network Canvas to provide an end-to-end workflow with intuitive interfaces to enable researchers to design and conduct network interviews.

IMPLEMENTATION

Network Canvas consists of three applications (Architect, Interviewer and Server). All applications are written in JavaScript and run on Windows, macOS and Linux; Interviewer also runs on Android and iOS.

GENERAL FEATURES

Network Canvas substantially reduces the complexity and technical knowledge required to collect network data via three point-and-click applications. The tool has wide applicability for measuring contact and social influence networks in epidemiology.

AVAILABILITY

Network Canvas is open source and freely available [networkcanvas.com] under the GNU General Public License 3.

Distinguishing Simple and Complex Contagion Processes on Networks

Contagion processes on networks, including disease spreading, information diffusion, or social behaviors propagation, can be modeled as simple contagion, i.e., as a contagion process involving one connection at a time, or as complex contagion, in which multiple interactions are needed for a contagion event. Empirical data on spreading processes, however, even when available, do not easily allow us to uncover which of these underlying contagion mechanisms is at work. We propose a strategy to discriminate between these mechanisms upon the observation of a single instance of a spreading process. The strategy is based on the observation of the order in which network nodes are infected, and on its correlations with their local topology: these correlations differ between processes of simple contagion, processes involving threshold mechanisms, and processes driven by group interactions (i.e., by "higher-order" mechanisms). Our results improve our understanding of contagion processes and provide a method using only limited information to distinguish between several possible contagion mechanisms.

Who Gets the Flu? Individualized Validation of Influenza-like Illness in Urban Spaces

Urban dwellers are exposed to communicable diseases, such as influenza, in various urban spaces. Current disease models are able to predict health outcomes at the individual scale but are mostly validated at coarse scales due to the lack of fine-scaled ground truth data. Further, a large number of transmission-driving factors have been considered in these models. Because of the lack of individual-scaled validations, the effectiveness of factors at their intended scale is not substantiated. These gaps significantly undermine the efficacy of the models in assessing the vulnerability of individuals, communities, and urban society. The objectives of this study are twofold. First, we aim to model and, most importantly, validate influenza-like illness (ILI) symptoms at the individual scale based on four sets of transmission-driving factors pertinent to home-work space, service space, ambient environment, and demographics. The effort is supported by an ensemble approach. For the second objective, we investigate the effectiveness of the factor sets through an impact analysis. The validation accuracy reaches 73.2-95.1%. The validation substantiates the effectiveness of factors pertinent to urban spaces and unveils the underlying mechanism that connects urban spaces and population health. With more fine-scaled health data becoming available, the findings of this study may see increasing value in informing policies that improve population health and urban livability.

Spatial immunization to abate disease spreading in transportation hubs

Proximity social interactions are crucial for infectious diseases transmission. Crowded agglomerations pose serious risk of triggering superspreading events. Locations like transportation hubs (airports and stations) are designed to optimize logistic efficiency, not to reduce crowding, and are characterized by a constant in and out flow of people. Here, we analyze the paradigmatic example of London Heathrow, one of the busiest European airports. Thanks to a dataset of anonymized individuals' trajectories, we can model the spreading of different diseases to localize the contagion hotspots and to propose a spatial immunization policy targeting them to reduce disease spreading risk. We also detect the most vulnerable destinations to contagions produced at the airport and quantify the benefits of the spatial immunization technique to prevent regional and global disease diffusion. This method is immediately generalizable to train, metro and bus stations and to other facilities such as commercial or convention centers.

Large-scale measurement of aggregate human colocation patterns for epidemiological modeling

To understand and model public health emergencies, epidemiologists need data that describes how humans are moving and interacting across physical space. Such data has traditionally been difficult for researchers to obtain with the temporal resolution and geographic breadth that is needed to study, for example, a global pandemic. This paper describes Colocation Maps, which are spatial network datasets that have been developed within Meta's Data For Good program. These Maps estimate how often people from different regions are colocated: in particular, for a pair of geographic regions x and y, these Maps estimate the rate at which a randomly chosen person from x and a randomly chosen person from y are simultaneously located in the same place during a randomly chosen minute in a given week. These datasets are well suited to parametrize metapopulation models of disease spread or to measure temporal changes in interactions between people from different regions; indeed, they have already been used for both of these purposes during the COVID-19 pandemic. In this paper, we show how Colocation Maps differ from existing data sources, describe how the datasets are built, provide examples of their use in compartmental modeling, and summarize ideas for further development of these and related datasets. Among the findings of this study, we observe that a pair of regions can exhibit high colocation despite few people moving between those regions. Additionally, for the purposes of clarifying how to interpret and utilize Colocation Maps, we scrutinize the Maps' built-in assumptions about representativeness and contact heterogeneity.

Impacts of observation frequency on proximity contact data and modeled transmission dynamics

Transmission of many communicable diseases depends on proximity contacts among humans. Modeling the dynamics of proximity contacts can help determine whether an outbreak is likely to trigger an epidemic. While the advent of commodity mobile devices has eased the collection of proximity contact data, battery capacity and associated costs impose tradeoffs between the observation frequency and scanning duration used for contact detection. The choice of observation frequency should depend on the characteristics of a particular pathogen and accompanying disease. We downsampled data from five contact network studies, each measuring participant-participant contact every 5 minutes for durations of four or more weeks. These studies included a total of 284 participants and exhibited different community structures. We found that for epidemiological models employing high-resolution proximity data, both the observation method and observation frequency configured to collect proximity data impact the simulation results. This impact is subject to the population's characteristics as well as pathogen infectiousness. By comparing the performance of two observation methods, we found that in most cases, half-hourly Bluetooth discovery for one minute can collect proximity data that allows agent-based transmission models to produce a reasonable estimation of the attack rate, but more frequent Bluetooth discovery is preferred to model individual infection risks or for highly transmissible pathogens. Our findings inform the empirical basis for guidelines to inform data collection that is both efficient and effective.

User-Centric Proximity Estimation Using Smartphone Radio Fingerprinting

The integration of infectious disease modeling with the data collection process is crucial to reach its maximum potential, and remains a significant research challenge. Ensuring a solid empirical foundation for models used to fill gaps in data and knowledge is of paramount importance. Personal wireless devices, such as smartphones, smartwatches and wireless bracelets, can serve as a means of bridging the gap between empirical data and the mathematical modeling of human contacts and networking. In this paper, we develop, implement, and evaluate concepts and architectures for advanced user-centric proximity estimation based on smartphone radio environment monitoring. We investigate innovative methods for the estimation of proximity, based on a person-radio-environment trace recorded by the smartphone, and define the proximity parameter. For this purpose, we developed a smartphone application and back-end services. The results show that, with the proposed procedure, we can estimate the proximity of two devices in terms of near, medium, and far distance with reasonable accuracy in real-world case scenarios.

Optimal vaccine roll-out strategies including social distancing for pandemics

Non-pharmacological interventions (NPIs), principally social distancing, in combination with effective vaccines, aspire to develop a protective immunity shield against pandemics and particularly against the COVID-19 pandemic. In this study, an agent-based network model with small-world topology is employed to find optimal policies against pandemics, including social distancing and vaccination strategies. The agents' states are characterized by a variation of the SEIR model (susceptible, exposed, infected, recovered). To explore optimal policies, an equation-free method is proposed to solve the inverse problem of calibrating an agent's infection rate with respect to the vaccination efficacy. The results show that prioritizing the first vaccine dose in combination with mild social restrictions, is sufficient to control the pandemic, with respect to the number of deaths. Moreover, for the same mild number of social contacts, we find an optimal vaccination ratio of 0.85 between older people of ages > 65 compared to younger ones.

Impact of contact data resolution on the evaluation of interventions in mathematical models of infectious diseases

Computational models offer a unique setting to test strategies to mitigate the spread of infectious diseases, providing useful insights to applied public health. To be actionable, models need to be informed by data, which can be available at different levels of detail. While high-resolution data describing contacts between individuals are increasingly available, data gathering remains challenging, especially during a health emergency. Many models thus use synthetic data or coarse information to evaluate intervention protocols. Here, we evaluate how the representation of contact data might affect the impact of various strategies in models, in the realm of COVID-19 transmission in educational and work contexts. Starting from high-resolution contact data, we use detailed to coarse data representations to inform a model of SARS-CoV-2 transmission and simulate different mitigation strategies. We find that coarse data representations estimate a lower risk of superspreading events. However, the rankings of protocols according to their efficiency or cost remain coherent across representations, ensuring the consistency of model findings to inform public health advice. Caution should be taken, however, on the quantitative estimations of those benefits and costs triggering the adoption of protocols, as these may depend on data representation.

Using high-resolution contact networks to evaluate SARS-CoV-2 transmission and control in large-scale multi-day events

The emergence of highly transmissible SARS-CoV-2 variants has created a need to reassess the risk posed by increasing social contacts as countries resume pre-pandemic activities, particularly in the context of resuming large-scale events over multiple days. To examine how social contacts formed in different activity settings influences interventions required to control Delta variant outbreaks, we collected high-resolution data on contacts among passengers and crew on cruise ships and combined the data with network transmission models. We found passengers had a median of 20 (IQR 10-36) unique close contacts per day, and over 60% of their contact episodes were made in dining or sports areas where mask wearing is typically limited. In simulated outbreaks, we found that vaccination coverage and rapid antigen tests had a larger effect than mask mandates alone, indicating the importance of combined interventions against Delta to reduce event risk in the vaccine era.

Challenges for modelling interventions for future pandemics

Mathematical modelling and statistical inference provide a framework to evaluate different non-pharmaceutical and pharmaceutical interventions for the control of epidemics that has been widely used during the COVID-19 pandemic. In this paper, lessons learned from this and previous epidemics are used to highlight the challenges for future pandemic control. We consider the availability and use of data, as well as the need for correct parameterisation and calibration for different model frameworks. We discuss challenges that arise in describing and distinguishing between different interventions, within different modelling structures, and allowing both within and between host dynamics. We also highlight challenges in modelling the health economic and political aspects of interventions. Given the diversity of these challenges, a broad variety of interdisciplinary expertise is needed to address them, combining mathematical knowledge with biological and social insights, and including health economics and communication skills. Addressing these challenges for the future requires strong cross-disciplinary collaboration together with close communication between scientists and policy makers.

Challenges in estimation, uncertainty quantification and elicitation for pandemic modelling

The estimation of parameters and model structure for informing infectious disease response has become a focal point of the recent pandemic. However, it has also highlighted a plethora of challenges remaining in the fast and robust extraction of information using data and models to help inform policy. In this paper, we identify and discuss four broad challenges in the estimation paradigm relating to infectious disease modelling, namely the Uncertainty Quantification framework, data challenges in estimation, model-based inference and prediction, and expert judgement. We also postulate priorities in estimation methodology to facilitate preparation for future pandemics.

Revealing mechanisms of infectious disease spread through empirical contact networks

The spread of pathogens fundamentally depends on the underlying contacts between individuals. Modeling the dynamics of infectious disease spread through contact networks, however, can be challenging due to limited knowledge of how an infectious disease spreads and its transmission rate. We developed a novel statistical tool, INoDS (Identifying contact Networks of infectious Disease Spread) that estimates the transmission rate of an infectious disease outbreak, establishes epidemiological relevance of a contact network in explaining the observed pattern of infectious disease spread and enables model comparison between different contact network hypotheses. We show that our tool is robust to incomplete data and can be easily applied to datasets where infection timings of individuals are unknown. We tested the reliability of INoDS using simulation experiments of disease spread on a synthetic contact network and find that it is robust to incomplete data and is reliable under different settings of network dynamics and disease contagiousness compared with previous approaches. We demonstrate the applicability of our method in two host-pathogen systems: Crithidia bombi in bumblebee colonies and Salmonella in wild Australian sleepy lizard populations. INoDS thus provides a novel and reliable statistical tool for identifying transmission pathways of infectious disease spread. In addition, application of INoDS extends to understanding the spread of novel or emerging infectious disease, an alternative approach to laboratory transmission experiments, and overcoming common data-collection constraints.

Network models are widely used to understand and predict infectious disease spread in human and animal populations. However, the choice of network model often relies on subjective expert knowledge or disease transmission experiments that are time-consuming and difficult to perform. We developed a novel tool, called INoDS (Identifying contact Networks of infectious Disease Spread), that uses robust statistical approach to establish relevance of a network model in explaining transmission pathways of an infectious disease outbreak. We used computer simulations and real-world dataset to test the accuracy of our tool and robustness to missing data. We found that INoDS is robust to common data-collection constraints, broadly applicable and accurate compared to current approaches. The tool that we have developed can therefore provide immediate epidemiological insights in the event of an epidemic outbreak, and can be used to improve targeted disease control.

Transmission dynamics of an antimicrobial resistant Campylobacter jejuni lineage in New Zealand's commercial poultry network

Understanding the relative contribution of different between-farm transmission pathways is essential in guiding recommendations for mitigating disease spread. This study investigated the association between contact pathways linking poultry farms in New Zealand and the genetic relatedness of antimicrobial resistant Campylobacter jejuni Sequence Type 6964 (ST-6964), with the aim of identifying the most likely contact pathways that contributed to its rapid spread across the industry. Whole-genome sequencing was performed on 167C. jejuni ST-6964 isolates sampled from across 30 New Zealand commercial poultry enterprises. The genetic relatedness between isolates was determined using whole genome multilocus sequence typing (wgMLST). Permutational multivariate analysis of variance and distance-based linear models were used to explore the strength of the relationship between pairwise genetic associations among the C. jejuni isolates and each of several pairwise distance matrices, indicating either the geographical distance between farms or the network distance of transportation vehicles. Overall, a significant association was found between the pairwise genetic relatedness of the C. jejuni isolates and the parent company, the road distance and the network distance of transporting feed vehicles. This result suggests that the transportation of feed within the commercial poultry industry as well as other local contacts between flocks, such as the movements of personnel, may have played a significant role in the spread of C. jejuni. However, further information on the historical contact patterns between farms is needed to fully characterise the risk of these pathways and to understand how they could be targeted to reduce the spread of C. jejuni.

Reconstructing contact network structure and cross-immunity patterns from multiple infection histories

Interactions within a population shape the spread of infectious diseases but contact patterns between individuals are difficult to access. We hypothesised that key properties of these patterns can be inferred from multiple infection data in longitudinal follow-ups. We developed a simulator for epidemics with multiple infections on networks and analysed the resulting individual infection time series by introducing similarity metrics between hosts based on their multiple infection histories. We find that, depending on infection multiplicity and network sampling, multiple infection summary statistics can recover network properties such as degree distribution. Furthermore, we show that by mining simulation outputs for multiple infection patterns, one can detect immunological interference between pathogens (i.e. the fact that past infections in a host condition future probability of infection). The combination of individual-based simulations and analysis of multiple infection histories opens promising perspectives to infer and validate transmission networks and immunological interference for infectious diseases from longitudinal cohort data.

Towards a unified generic framework to define and observe contacts between livestock and wildlife: a systematic review

Wild animals are the source of many pathogens of livestock and humans. Concerns about the potential transmission of economically important and zoonotic diseases from wildlife have led to increased surveillance at the livestock-wildlife interface. Knowledge of the types, frequency and duration of contacts between livestock and wildlife is necessary to identify risk factors for disease transmission and to design possible mitigation strategies. Observing the behaviour of many wildlife species is challenging due to their cryptic nature and avoidance of humans, meaning there are relatively few studies in this area. Further, a consensus on the definition of what constitutes a 'contact' between wildlife and livestock is lacking. A systematic review was conducted to investigate which livestock-wildlife contacts have been studied and why, as well as the methods used to observe each species. Over 30,000 publications were screened, of which 122 fulfilled specific criteria for inclusion in the analysis. The majority of studies examined cattle contacts with badgers or with deer; studies involving wild pig contacts with cattle or with domestic pigs were the next most frequent. There was a range of observational methods including motion-activated cameras and global positioning system collars. As a result of the wide variation and lack of consensus in the definitions of direct and indirect contacts, we developed a unified framework to define livestock-wildlife contacts that is sufficiently flexible to be applied to most wildlife and livestock species for non-vector-borne diseases. We hope this framework will help standardise the collection and reporting of contact data; a valuable step towards being able to compare the efficacy of wildlife-livestock observation methods. In doing so, it may aid the development of better disease transmission models and improve the design and effectiveness of interventions to reduce or prevent disease transmission.

Patch dynamics modeling framework from pathogens' perspective: Unified and standardized approach for complicated epidemic systems

Mathematical models are powerful tools to investigate, simulate, and evaluate potential interventions for infectious diseases dynamics. Much effort has focused on the Susceptible-Infected-Recovered (SIR)-type compartment models. These models consider host populations and measure change of each compartment. In this study, we propose an alternative patch dynamic modeling framework from pathogens' perspective. Each patch, the basic module of this modeling framework, has four standard mechanisms of pathogen population size change: birth (replication), death, inflow, and outflow. This framework naturally distinguishes between-host transmission process (inflow and outflow) and within-host infection process (replication) during the entire transmission-infection cycle. We demonstrate that the SIR-type model is actually a special cross-sectional and discretized case of our patch dynamics model in pathogens' viewpoint. In addition, this patch dynamics modeling framework is also an agent-based model from hosts' perspective by incorporating individual host's specific traits. We provide an operational standard to formulate this modular-designed patch dynamics model. Model parameterization is feasible with a wide range of sources, including genomics data, surveillance data, electronic health record, and from other emerging technologies such as multiomics. We then provide two proof-of-concept case studies to tackle some of the existing challenges of SIR-type models: sexually transmitted disease and healthcare acquired infections. This patch dynamics modeling framework not only provides theoretical explanations to known phenomena, but also generates novel insights of disease dynamics from a more holistic viewpoint. It is also able to simulate and handle more complicated scenarios across biological scales such as the current COVID-19 pandemic.

Adequacy of SEIR models when epidemics have spatial structure: Ebola in Sierra Leone

Dynamic SEIR (Susceptible, Exposed, Infectious, Removed) compartmental models provide a tool for predicting the size and duration of both unfettered and managed outbreaks-the latter in the context of interventions such as case detection, patient isolation, vaccination and treatment. The reliability of this tool depends on the validity of key assumptions that include homogeneity of individuals and spatio-temporal homogeneity. Although the SEIR compartmental framework can easily be extended to include demographic (e.g. age) and additional disease (e.g. healthcare workers) classes, dependence of transmission rates on time, and metapopulation structure, fitting such extended models is hampered by both a proliferation of free parameters and insufficient or inappropriate data. This raises the question of how effective a tool the basic SEIR framework may actually be. We go some way here to answering this question in the context of the 2014-2015 outbreak of Ebola in West Africa by comparing fits of an SEIR time-dependent transmission model to both country- and district-level weekly incidence data. Our novel approach in estimating the effective-size-of-the-populations-at-risk ( N_eff) and initial number of exposed individuals ( E₀) at both district and country levels, as well as the transmission function parameters, including a time-to-halving-the-force-of-infection ( t_f/2) parameter, provides new insights into this Ebola outbreak. It reveals that the estimate R₀ ≈ 1.7 from country-level data appears to seriously underestimate R₀ ≈ 3.3 - 4.3 obtained from more spatially homogeneous district-level data. Country-level data also overestimate t_f/2 ≈ 22 weeks, compared with 8-10 weeks from district-level data. Additionally, estimates for the duration of individual infectiousness is around two weeks from spatially inhomogeneous country-level data compared with 2.4-4.5 weeks from spatially more homogeneous district-level data, which estimates are rather high compared with most values reported in the literature. This article is part of the theme issue 'Modelling infectious disease outbreaks in humans, animals and plants: approaches and important themes'. This issue is linked with the subsequent theme issue 'Modelling infectious disease outbreaks in humans, animals and plants: epidemic forecasting and control'.

A Probabilistic Infection Model for Efficient Trace-Prediction of Disease Outbreaks in Contact Networks

We propose a novel method which we call the Probabilistic Infection Model (PIM). Instead of stochastically assigning exactly one state to each agent at a time, PIM tracks the likelihood of each agent being in a particular state. Thus, a particular agent can exist in multiple disease states concurrently. Our model gives an improved resolution of transitions between states, and allows for a more comprehensive view of outbreak dynamics at the individual level. Moreover, by using a probabilistic approach, our model gives a representative understanding of the overall trajectories of simulated outbreaks without the need for numerous (order of hundreds) of repeated Monte Carlo simulations.

We simulate our model over a contact network constructed using registration data of university students. We model three diseases; measles and two strains of influenza. We compare the results obtained by PIM with those obtained by simulating stochastic SEIR models over the same the contact network. The results demonstrate that the PIM can successfully replicate the averaged results from numerous simulations of a stochastic model in a single deterministic simulation.

Incorporating Contact Network Uncertainty in Individual Level Models of Infectious Disease using Approximate Bayesian Computation

Infectious disease transmission between individuals in a heterogeneous population is often best modelled through a contact network. However, such contact network data are often unobserved. Such missing data can be accounted for in a Bayesian data augmented framework using Markov chain Monte Carlo (MCMC). Unfortunately, fitting models in such a framework can be highly computationally intensive. We investigate the fitting of network-based infectious disease models with completely unknown contact networks using approximate Bayesian computation population Monte Carlo (ABC-PMC) methods. This is done in the context of both simulated data, and data from the UK 2001 foot-and-mouth disease epidemic. We show that ABC-PMC is able to obtain reasonable approximations of the underlying infectious disease model with huge savings in computation time when compared to a full Bayesian MCMC analysis.

Incorporating genomic methods into contact networks to reveal new insights into animal behavior and infectious disease dynamics

Utilization of contact networks has provided opportunities for assessing the dynamic interplay between pathogen transmission and host behavior. Genomic techniques have, in their own right, provided new insight into complex questions in disease ecology, and the increasing accessibility of genomic approaches means more researchers may seek out these tools. The integration of network and genomic approaches provides opportunities to examine the interaction between behavior and pathogen transmission in new ways and with greater resolution. While a number of studies have begun to incorporate both contact network and genomic approaches, a great deal of work has yet to be done to better integrate these techniques. In this review, we give a broad overview of how network and genomic approaches have each been used to address questions regarding the interaction of social behavior and infectious disease, and then discuss current work and future horizons for the merging of these techniques.

A systematic review of scabies transmission models and data to evaluate the cost-effectiveness of scabies interventions

Background

Scabies is a common dermatological condition, affecting more than 130 million people at any time. To evaluate and/or predict the effectiveness and cost-effectiveness of scabies interventions, disease transmission modelling can be used.

Objective

To review published scabies models and data to inform the design of a comprehensive scabies transmission modelling framework to evaluate the cost-effectiveness of scabies interventions.

Methods

Systematic literature search in PubMed, Medline, Embase, CINAHL, and the Cochrane Library identified scabies studies published since the year 2000. Selected papers included modelling studies and studies on the life cycle of scabies mites, patient quality of life and resource use. Reference lists of reviews were used to identify any papers missed through the search strategy. Strengths and limitations of identified scabies models were evaluated and used to design a modelling framework. Potential model inputs were identified and discussed.

Findings

Four scabies models were published: a Markov decision tree, two compartmental models, and an agent-based, network-dependent Monte Carlo model. None of the models specifically addressed crusted scabies, which is associated with high morbidity, mortality, and increased transmission. There is a lack of reliable, comprehensive information about scabies biology and the impact this disease has on patients and society.

Discussion

Clinicians and health economists working in the field of scabies are encouraged to use the current review to inform disease transmission modelling and economic evaluations on interventions against scabies.

Scabies is a neglected tropical disease affecting more than 130 million people, with major costs on health care systems worldwide. While effective treatments exist, it is unknown which treatment strategies result in the best outcomes against the lowest costs, and to what extent this differs between communities. Health economic modelling can help answer these questions, but has rarely been used in this disease area. This review discusses all available scabies transmission models (n = 4), and uses them to create a new, comprehensive modelling framework. This framework can be used as aid for creating a scabies transmission model, the details of which will be determined by the context (population) and the question being addressed. The current paper also reviews the data that is needed to inform scabies modelling: on scabies biology, quality of life and resource use. Unfortunately, available data is limited and particularly data on crusted scabies (associated with high morbidity and mortality rates) is rare. With this review, we hope to assist researchers and policy makers to predict and/or evaluate the cost-effectiveness of interventions against scabies in their population(s) of interest. To tackle scabies, it is key to use effective treatment strategies in a cost-effective and sustainable way. The models and data described in this review, may help researchers, clinicians and funding bodies to facilitate this.

Transmission on empirical dynamic contact networks is influenced by data processing decisions

Dynamic contact data can be used to inform disease transmission models, providing insight into the dynamics of infectious diseases. Such data often requires extensive processing for use in models or analysis. Therefore, processing decisions can potentially influence the topology of the contact network and the simulated disease transmission dynamics on the network. In this study, we examine how four processing decisions, including temporal sampling window (TSW), spatial threshold of contact (SpTh), minimum contact duration (MCD), and temporal aggregation (daily or hourly) influence the information content of contact data (indicated by changes in entropy) as well as disease transmission model dynamics. We found that changes made to information content by processing decisions translated to significant impacts to the transmission dynamics of disease models using the contact data. In particular, we found that SpTh had the largest independent influence on information content, and that some output metrics (R₀, time to peak infection) were more sensitive to changes in information than others (epidemic extent). These findings suggest that insights gained from transmission modeling using dynamic contact data can be influenced by processing decisions alone, emphasizing the need to carefully consideration them prior to using contact-based models to conduct analyses, compare different datasets, or inform policy decisions.

Phylogenies from dynamic networks

The relationship between the underlying contact network over which a pathogen spreads and the pathogen phylogenetic trees that are obtained presents an opportunity to use sequence data to learn about contact networks that are difficult to study empirically. However, this relationship is not explicitly known and is usually studied in simulations, often with the simplifying assumption that the contact network is static in time, though human contact networks are dynamic. We simulate pathogen phylogenetic trees on dynamic Erdős-Renyi random networks and on two dynamic networks with skewed degree distribution, of which one is additionally clustered. We use tree shape features to explore how adding dynamics changes the relationships between the overall network structure and phylogenies. Our tree features include the number of small substructures (cherries, pitchforks) in the trees, measures of tree imbalance (Sackin index, Colless index), features derived from network science (diameter, closeness), as well as features using the internal branch lengths from the tip to the root. Using principal component analysis we find that the network dynamics influence the shapes of phylogenies, as does the network type. We also compare dynamic and time-integrated static networks. We find, in particular, that static network models like the widely used Barabasi-Albert model can be poor approximations for dynamic networks. We explore the effects of mis-specifying the network on the performance of classifiers trained identify the transmission rate (using supervised learning methods). We find that both mis-specification of the underlying network and its parameters (mean degree, turnover rate) have a strong adverse effect on the ability to estimate the transmission parameter. We illustrate these results by classifying HIV trees with a classifier that we trained on simulated trees from different networks, infection rates and turnover rates. Our results point to the importance of correctly estimating and modelling contact networks with dynamics when using phylodynamic tools to estimate epidemiological parameters.

Structure and consistency of self-reported social contact networks in British secondary schools

Self-reported social mixing patterns are commonly used in mathematical models of infectious diseases. It is particularly important to quantify patterns for school-age children given their disproportionate role in transmission, but it remains unclear how the structure of such social interactions changes over time. By integrating data collection into a public engagement programme, we examined self-reported contact networks in year 7 groups in four UK secondary schools. We collected data from 460 unique participants across four rounds of data collection conducted between January and June 2015, with 7,315 identifiable contacts reported in total. Although individual-level contacts varied over the study period, we were able to obtain out-of-sample accuracies of more than 90% and F-scores of 0.49-0.84 when predicting the presence or absence of social contacts between specific individuals across rounds of data collection. Network properties such as clustering and number of communities were broadly consistent within schools between survey rounds, but varied significantly between schools. Networks were assortative according to gender, and to a lesser extent school class, with the estimated clustering coefficient larger among males in all surveyed co-educational schools. Our results demonstrate that it is feasible to collect longitudinal self-reported social contact data from school children and that key properties of these data are consistent between rounds of data collection.

Epidemics on dynamic networks

In many populations, the patterns of potentially infectious contacts are transients that can be described as a network with dynamic links. The relative timescales of link and contagion dynamics and the characteristics that drive their tempos can lead to important differences to the static case. Here, we propose some essential nomenclature for their analysis, and then review the relevant literature. We describe recent advances in they apply to infection processes, considering all of the methods used to record, measure and analyse them, and their implications for disease transmission. Finally, we outline some key challenges and opportunities in the field.

Temporally Varying Relative Risks for Infectious Diseases: Implications for Infectious Disease Control

Risks for disease in some population groups relative to others (relative risks) are usually considered to be consistent over time, although they are often modified by other, nontemporal factors. For infectious diseases, in which overall incidence often varies substantially over time, the patterns of temporal changes in relative risks can inform our understanding of basic epidemiologic questions. For example, recent studies suggest that temporal changes in relative risks of infection over the course of an epidemic cycle can both be used to identify population groups that drive infectious disease outbreaks, and help elucidate differences in the effect of vaccination against infection (that is relevant to transmission control) compared with its effect against disease episodes (that reflects individual protection). Patterns of change in the age groups affected over the course of seasonal outbreaks can provide clues to the types of pathogens that could be responsible for diseases for which an infectious cause is suspected. Changing apparent efficacy of vaccines during trials may provide clues to the vaccine's mode of action and/or indicate risk heterogeneity in the trial population. Declining importance of unusual behavioral risk factors may be a signal of increased local transmission of an infection. We review these developments and the related public health implications.

The role of venues in structuring HIV, sexually transmitted infections, and risk networks among men who have sex with men

Background

Venues form part of the sampling frame for time-location sampling, an approach often used for HIV surveillance. While sampling location is often regarded as a nuisance factor, venues may play a central role in structuring risk networks. We investigated individual reports of risk behaviors and infections among men who have sex with men (MSM) attending different venues to examine structuring of HIV risk behaviors. However, teasing apart ‘risky people’ from ‘risky places’ is difficult, as individuals cannot be randomized to attend different venues. However, we can emulate this statistically using marginal structural models, which inversely weight individuals according to their estimated probability of attending the venue.

Methods

We conducted a cross-sectional survey of 609 MSM patrons of 14 bars in San Diego, California, recruited using the Priorities for Local AIDS Control Efforts (PLACE) methodology, which consists of a multi-level identification and assessment of venues for HIV risk through population surveys.

Results and discussion

Venues differed by many factors, including participants’ reported age, ethnicity, number of lifetime male partners, past sexually transmitted infection (STI), and HIV status. In multivariable marginal structural models, venues demonstrated structuring of HIV+ status, past STI, and methamphetamine use, independently of individual-level characteristics.

Conclusions

Studies using time-location sampling should consider venue as an important covariate, and the use of marginal structural models may help to identify risky venues. This may assist in widespread, economically feasible and sustainable targeted surveillance and prevention. A more mechanistic understanding of how ‘risky venues’ emerge and structure risk is needed.

Electronic supplementary material

The online version of this article (10.1186/s12889-018-5140-3) contains supplementary material, which is available to authorized users.

Temporal variation of human encounters and the number of locations in which they occur: a longitudinal study of Hong Kong residents

Patterns of social contact between individuals are important for the transmission of many pathogens and shaping patterns of immunity at the population scale. To refine our understanding of how human social behaviour may change over time, we conducted a longitudinal study of Hong Kong residents. We recorded the social contact patterns for 1450 individuals, up to four times each between May 2012 and September 2013. We found individuals made contact with an average of 12.5 people within 2.9 geographical locations, and spent an average estimated total duration of 9.1 h in contact with others during a day. Distributions of the number of contacts and locations in which contacts were made were not significantly different between study waves. Encounters were assortative by age, and the age mixing pattern was broadly consistent across study waves. Fitting regression models, we examined the association of contact rates (number of contacts, total duration of contact, number of locations) with covariates and calculated the inter- and intra-participant variation in contact rates. Participant age was significantly associated with the number of contacts made, the total duration of contact and the number of locations in which contact occurred, with children and parental-age adults having the highest rates of contact. The number of contacts and contact duration increased with the number of contact locations. Intra-individual variation in contact rate was consistently greater than inter-individual variation. Despite substantial individual-level variation, remarkable consistency was observed in contact mixing at the population scale. This suggests that aggregate measures of mixing behaviour derived from cross-sectional information may be appropriate for population-scale modelling purposes, and that if more detailed models of social interactions are required for improved public health modelling, further studies are needed to understand the social processes driving intra-individual variation.

Evaluating empirical contact networks as potential transmission pathways for infectious diseases

Networks are often used to incorporate heterogeneity in contact patterns in mathematical models of pathogen spread. However, few tools exist to evaluate whether potential transmission pathways in a population are adequately represented by an observed contact network. Here, we describe a novel permutation-based approach, the network k-test, to determine whether the pattern of cases within the observed contact network are likely to have resulted from transmission processes in the network, indicating that the network represents potential transmission pathways between nodes. Using simulated data of pathogen spread, we compare the power of this approach to other commonly used analytical methods. We test the robustness of this technique across common sampling constraints, including undetected cases, unobserved individuals and missing interaction data. We also demonstrate the application of this technique in two case studies of livestock and wildlife networks. We show that the power of the k-test to correctly identify the epidemiologic relevance of contact networks is substantially greater than other methods, even when 50% of contact or case data are missing. We further demonstrate that the impact of missing data on network analysis depends on the structure of the network and the type of missing data.

Estimating the epidemic risk using non-uniformly sampled contact data

Many datasets describing contacts in a population suffer from incompleteness due to population sampling and underreporting of contacts. Data-driven simulations of spreading processes using such incomplete data lead to an underestimation of the epidemic risk, and it is therefore important to devise methods to correct this bias. We focus here on a non-uniform sampling of the contacts between individuals, aimed at mimicking the results of diaries or surveys, and consider as case studies two datasets collected in different contexts. We show that using surrogate data built using a method developed in the case of uniform population sampling yields an improvement with respect to the use of the sampled data but is strongly limited by the underestimation of the link density in the sampled network. We put forward a second method to build surrogate data that assumes knowledge of the density of links within one of the groups forming the population. We show that it gives very good results when the population is strongly structured, and discuss its limitations in the case of a population with a weaker group structure. These limitations highlight the interest of measurements using wearable sensors able to yield accurate information on the structure and durations of contacts.

The Timing of Pertussis Cases in Unvaccinated Children in an Outbreak Year: Oregon 2012

OBJECTIVE

To assess whether, during a 2012 pertussis outbreak, unvaccinated and poorly vaccinated cases occurred earlier on a community level.

STUDY DESIGN

Pediatric pertussis among children 2 months to 10 years of age in the Oregon Sentinel Surveillance region during an epidemic starting at the beginning of 2012 were stratified by immunization status, age, zip code, and calendar date of disease onset. Differences in median onset as days between fully or mostly vaccinated, poorly vaccinated, and unvaccinated cases were examined overall and within local zip code areas. Disease clusters also were examined using SatScan analysis.

RESULTS

Overall, 351 pertussis cases occurred among children aged 2 months to 10 years of age residing in 72 distinct zipcodes. Among unvaccinated or poorly vaccinated cases, their median date of onset was at calendar day 117 (April 26, 2012), whereas for those who were fully or mostly vaccinated the median date of onset was 41 days later, at day 158 (June 6, 2012). Within each local zip code area, the unvaccinated cases were 3.2 times more likely than vaccinated cases to have earlier median dates of onset (95% CI 2.9-3.6).

CONCLUSION

In this outbreak, pertussis cases among unvaccinated children represented an earlier spread of disease across local areas. Controlling outbreaks may require attention to the composition and location of the unvaccinated.

Phylodynamics on local sexual contact networks

Phylodynamic models are widely used in infectious disease epidemiology to infer the dynamics and structure of pathogen populations. However, these models generally assume that individual hosts contact one another at random, ignoring the fact that many pathogens spread through highly structured contact networks. We present a new framework for phylodynamics on local contact networks based on pairwise epidemiological models that track the status of pairs of nodes in the network rather than just individuals. Shifting our focus from individuals to pairs leads naturally to coalescent models that describe how lineages move through networks and the rate at which lineages coalesce. These pairwise coalescent models not only consider how network structure directly shapes pathogen phylogenies, but also how the relationship between phylogenies and contact networks changes depending on epidemic dynamics and the fraction of infected hosts sampled. By considering pathogen phylogenies in a probabilistic framework, these coalescent models can also be used to estimate the statistical properties of contact networks directly from phylogenies using likelihood-based inference. We use this framework to explore how much information phylogenies retain about the underlying structure of contact networks and to infer the structure of a sexual contact network underlying a large HIV-1 sub-epidemic in Switzerland.

Phylodynamic models relate the branching pattern of a pathogen’s phylogenetic tree to the tree-like growth of an epidemic as it spreads through a host population. Such models are increasingly used to learn about the epidemiology of different pathogens. We extend current models to consider the structure of host contact networks—the web of physical interactions through which pathogens spread. By considering how local interactions among hosts shape the phylogeny of a pathogen, our models offer a “pathogen’s eye view” of these networks. Our models also provide a statistical framework that can be used to infer network structure directly from phylogenies, which we use to estimate the properties of a sexual contact network in Switzerland from a HIV phylogeny.

Estimating contact patterns relevant to the spread of infectious diseases in Russia

Understanding human mixing patterns is the key to provide public health decision makers with model-based evaluation of strategies for the control of infectious diseases. Here we conducted a population-based survey in Tomsk, Russia, asking participants to record all their contacts in physical person during the day. We estimated 9.8 contacts per person per day on average, 15.2 when including additional estimated professional contacts. We found that contacts were highly assortative by age, especially for school-age individuals, and the number of contacts negatively correlated with the age of the participant. The network of contacts was quite clustered, with the majority of contacts (about 72%) occurring between family members, students of the same school/university, and work colleagues. School represents the location where the largest number of contacts was recorded - students contacted about 7 individuals per day at school. Our modeling analysis based on the recorded contact patterns supports the importance of modeling age-mixing patterns - we show that, in the case of an epidemic caused by a novel influenza virus, school-age individuals would be the most affected age group, followed by adults aged 35-44 years. In conclusion, this study reveals an age-mixing pattern in general agreement with that estimated for European countries, although with several quantitative differences. The observed differences can be attributable to sociodemographic and cultural differences between countries. The age- and setting-specific contact matrices provided in this study could be instrumental for the design of control measures for airborne infections, specifically targeted on the characteristics of the Russian population.

Reconstructing contact network parameters from viral phylogenies

Models of the spread of disease in a population often make the simplifying assumption that the population is homogeneously mixed, or is divided into homogeneously mixed compartments. However, human populations have complex structures formed by social contacts, which can have a significant influence on the rate of epidemic spread. Contact network models capture this structure by explicitly representing each contact which could possibly lead to a transmission. We developed a method based on approximate Bayesian computation (ABC), a likelihood-free inference strategy, for estimating structural parameters of the contact network underlying an observed viral phylogeny. The method combines adaptive sequential Monte Carlo for ABC, Gillespie simulation for propagating epidemics though networks, and a kernel-based tree similarity score. We used the method to fit the Barabási-Albert network model to simulated transmission trees, and also applied it to viral phylogenies estimated from ten published HIV sequence datasets. This model incorporates a feature called preferential attachment (PA), whereby individuals with more existing contacts accumulate new contacts at a higher rate. On simulated data, we found that the strength of PA and the number of infected nodes in the network can often be accurately estimated. On the other hand, the mean degree of the network, as well as the total number of nodes, was not estimable with ABC. We observed sub-linear PA power in all datasets, as well as higher PA power in networks of injection drug users. These results underscore the importance of considering contact structures when performing phylodynamic inference. Our method offers the potential to quantitatively investigate the contact network structure underlying viral epidemics.

Contact diaries versus wearable proximity sensors in measuring contact patterns at a conference: method comparison and participants' attitudes

BACKGROUND

Studies measuring contact networks have helped to improve our understanding of infectious disease transmission. However, several methodological issues are still unresolved, such as which method of contact measurement is the most valid. Further, complete network analysis requires data from most, ideally all, members of a network and, to achieve this, acceptance of the measurement method. We aimed at investigating measurement error by comparing two methods of contact measurement - paper diaries vs. wearable proximity sensors - that were applied concurrently to the same population, and we measured acceptability.

METHODS

We investigated the contact network of one day of an epidemiology conference in September 2014. Seventy-six participants wore proximity sensors throughout the day while concurrently recording their contacts with other study participants in a paper-diary; they also reported on method acceptability.

RESULTS

There were 329 contact reports in the paper diaries, corresponding to 199 contacts, of which 130 were noted by both parties. The sensors recorded 316 contacts, which would have resulted in 632 contact reports if there had been perfect concordance in recording. We estimated the probabilities that a contact was reported in a diary as: P = 72 % for <5 min contact duration (significantly lower than the following, p < 0.05), P = 86 % for 5-15 min, P = 89 % for 15-60 min, and P = 94 % for >60 min. The sets of sensor-measured and self-reported contacts had a large intersection, but neither was a subset of the other. Participants' aggregated contact duration was mostly substantially longer in the diary data than in the sensor data. Twenty percent of respondents (>1 reported contact) stated that filling in the diary was too much work, 25 % of respondents reported difficulties in remembering contacts, and 93 % were comfortable having their conference contacts measured by sensors.

CONCLUSION

Reporting and recording were not complete; reporting was particularly incomplete for contacts <5 min. The types of contact that both methods are capable of detecting are partly different. Participants appear to have overestimated the duration of their contacts. Conducting a study with diaries or wearable sensors was acceptable to and mostly easily done by participants. Both methods can be applied meaningfully if their specific limitations are considered and incompleteness is accounted for.

Contact diaries versus wearable proximity sensors in measuring contact patterns at a conference: method comparison and participants' attitudes

Background

Studies measuring contact networks have helped to improve our understanding of infectious disease transmission. However, several methodological issues are still unresolved, such as which method of contact measurement is the most valid. Further, complete network analysis requires data from most, ideally all, members of a network and, to achieve this, acceptance of the measurement method. We aimed at investigating measurement error by comparing two methods of contact measurement – paper diaries vs. wearable proximity sensors – that were applied concurrently to the same population, and we measured acceptability.

Methods

We investigated the contact network of one day of an epidemiology conference in September 2014. Seventy-six participants wore proximity sensors throughout the day while concurrently recording their contacts with other study participants in a paper-diary; they also reported on method acceptability.

Results

There were 329 contact reports in the paper diaries, corresponding to 199 contacts, of which 130 were noted by both parties. The sensors recorded 316 contacts, which would have resulted in 632 contact reports if there had been perfect concordance in recording. We estimated the probabilities that a contact was reported in a diary as: P = 72 % for <5 min contact duration (significantly lower than the following, p < 0.05), P = 86 % for 5-15 min, P = 89 % for 15-60 min, and P = 94 % for >60 min. The sets of sensor-measured and self-reported contacts had a large intersection, but neither was a subset of the other. Participants’ aggregated contact duration was mostly substantially longer in the diary data than in the sensor data. Twenty percent of respondents (>1 reported contact) stated that filling in the diary was too much work, 25 % of respondents reported difficulties in remembering contacts, and 93 % were comfortable having their conference contacts measured by sensors.

Conclusion

Reporting and recording were not complete; reporting was particularly incomplete for contacts <5 min. The types of contact that both methods are capable of detecting are partly different. Participants appear to have overestimated the duration of their contacts. Conducting a study with diaries or wearable sensors was acceptable to and mostly easily done by participants. Both methods can be applied meaningfully if their specific limitations are considered and incompleteness is accounted for.

Electronic supplementary material

The online version of this article (doi:10.1186/s12879-016-1676-y) contains supplementary material, which is available to authorized users.

Distinction and connection between contact network, social network, and disease transmission network

In this paper we discuss the distinction and connection between three closely related networks in animal ecology and epidemiology studies: the contact, social, and disease transmission networks. We provide a robust theoretical definition and interpretation of these three networks, demonstrate that social and disease transmission networks can be derived as spanning subgraphs of contact network, and show examples based on real-world high-resolution cattle contact structure data. Furthermore, we establish a modeling framework to track potential disease transmission dynamics and construct transmission network based on the observed animal contact network.

How to Estimate Epidemic Risk from Incomplete Contact Diaries Data?

Social interactions shape the patterns of spreading processes in a population. Techniques such as diaries or proximity sensors allow to collect data about encounters and to build networks of contacts between individuals. The contact networks obtained from these different techniques are however quantitatively different. Here, we first show how these discrepancies affect the prediction of the epidemic risk when these data are fed to numerical models of epidemic spread: low participation rate, under-reporting of contacts and overestimation of contact durations in contact diaries with respect to sensor data determine indeed important differences in the outcomes of the corresponding simulations with for instance an enhanced sensitivity to initial conditions. Most importantly, we investigate if and how information gathered from contact diaries can be used in such simulations in order to yield an accurate description of the epidemic risk, assuming that data from sensors represent the ground truth. The contact networks built from contact sensors and diaries present indeed several structural similarities: this suggests the possibility to construct, using only the contact diary network information, a surrogate contact network such that simulations using this surrogate network give the same estimation of the epidemic risk as simulations using the contact sensor network. We present and compare several methods to build such surrogate data, and show that it is indeed possible to obtain a good agreement between the outcomes of simulations using surrogate and sensor data, as long as the contact diary information is complemented by publicly available data describing the heterogeneity of the durations of human contacts.

Schools, offices, hospitals play an important role in the spreading of epidemics. Information about interactions between individuals in such contexts can help understand the patterns of transmission and design ad hoc immunization strategies. Data about contacts can be collected through various techniques such as diaries or proximity sensors. Here, we first ask if the corresponding datasets give similar predictions of the epidemic risk when they are used to build a network of contacts among individuals. Not surprisingly, the answer is negative: indeed, if we consider data from sensors as the ground truth, diaries are affected by low participation rate, underreporting and overestimation of durations. Is it however possible, despite these biases, to use data from contact diaries to obtain sensible epidemic risk predictions? We show here that, thanks to the structural similarities between the two networks, it is possible to use the contact diaries to build surrogate versions of the contact network obtained from the sensor data, such that both yield the same epidemic risk estimation. We show that the construction of such surrogate networks can be performed using solely the information contained in the contact diaries, complemented by publicly available data on the heterogeneity of cumulative contact durations between individuals.

Quantifying social contacts in a household setting of rural Kenya using wearable proximity sensors

Close proximity interactions between individuals influence how infections spread. Quantifying close contacts in developing world settings, where such data is sparse yet disease burden is high, can provide insights into the design of intervention strategies such as vaccination. Recent technological advances have enabled collection of time-resolved face-to-face human contact data using radio frequency proximity sensors. The acceptability and practicalities of using proximity devices within the developing country setting have not been investigated. We present and analyse data arising from a prospective study of 5 households in rural Kenya, followed through 3 consecutive days. Pre-study focus group discussions with key community groups were held. All residents of selected households carried wearable proximity sensors to collect data on their close (<1.5 metres) interactions. Data collection for residents of three of the 5 households was contemporaneous. Contact matrices and temporal networks for 75 individuals are defined and mixing patterns by age and time of day in household contacts determined. Our study demonstrates the stability of numbers and durations of contacts across days. The contact durations followed a broad distribution consistent with data from other settings. Contacts within households occur mainly among children and between children and adults, and are characterised by daily regular peaks in the morning, midday and evening. Inter-household contacts are between adults and more sporadic when measured over several days. Community feedback indicated privacy as a major concern especially regarding perceptions of non-participants, and that community acceptability required thorough explanation of study tools and procedures. Our results show for a low resource setting how wearable proximity sensors can be used to objectively collect high-resolution temporal data without direct supervision. The methodology appears acceptable in this population following adequate community engagement on study procedures. A target for future investigation is to determine the difference in contact networks within versus between households. We suggest that the results from this study may be used in the design of future studies using similar electronic devices targeting communities, including households and schools, in the developing world context.

ELECTRONIC SUPPLEMENTARY MATERIAL

The online version of this article (doi:10.1140/epjds/s13688-016-0084-2) contains supplementary material.

Estimates of Social Contact in a Middle School Based on Self-Report and Wireless Sensor Data

Estimates of contact among children, used for infectious disease transmission models and understanding social patterns, historically rely on self-report logs. Recently, wireless sensor technology has enabled objective measurement of proximal contact and comparison of data from the two methods. These are mostly small-scale studies, and knowledge gaps remain in understanding contact and mixing patterns and also in the advantages and disadvantages of data collection methods. We collected contact data from a middle school, with 7th and 8th grades, for one day using self-report contact logs and wireless sensors. The data were linked for students with unique initials, gender, and grade within the school. This paper presents the results of a comparison of two approaches to characterize school contact networks, wireless proximity sensors and self-report logs. Accounting for incomplete capture and lack of participation, we estimate that "sensor-detectable", proximal contacts longer than 20 seconds during lunch and class-time occurred at 2 fold higher frequency than "self-reportable" talk/touch contacts. Overall, 55% of estimated talk-touch contacts were also sensor-detectable whereas only 15% of estimated sensor-detectable contacts were also talk-touch. Contacts detected by sensors and also in self-report logs had longer mean duration than contacts detected only by sensors (6.3 vs 2.4 minutes). During both lunch and class-time, sensor-detectable contacts demonstrated substantially less gender and grade assortativity than talk-touch contacts. Hallway contacts, which were ascertainable only by proximity sensors, were characterized by extremely high degree and short duration. We conclude that the use of wireless sensors and self-report logs provide complementary insight on in-school mixing patterns and contact frequency.

Epidemic risk from friendship network data: an equivalence with a non-uniform sampling of contact networks

Contacts between individuals play an important role in determining how infectious diseases spread. Various methods to gather data on such contacts co-exist, from surveys to wearable sensors. Comparisons of data obtained by different methods in the same context are however scarce, in particular with respect to their use in data-driven models of spreading processes. Here, we use a combined data set describing contacts registered by sensors and friendship relations in the same population to address this issue in a case study. We investigate if the use of the friendship network is equivalent to a sampling procedure performed on the sensor contact network with respect to the outcome of simulations of spreading processes: such an equivalence might indeed give hints on ways to compensate for the incompleteness of contact data deduced from surveys. We show that this is indeed the case for these data, for a specifically designed sampling procedure, in which respondents report their neighbors with a probability depending on their contact time. We study the impact of this specific sampling procedure on several data sets, discuss limitations of our approach and its possible applications in the use of data sets of various origins in data-driven simulations of epidemic processes.

Tracking social contact networks with online respondent-driven detection: who recruits whom?

Background

Transmission of respiratory pathogens in a population depends on the contact network patterns of individuals. To accurately understand and explain epidemic behaviour information on contact networks is required, but only limited empirical data is available. Online respondent-driven detection can provide relevant epidemiological data on numbers of contact persons and dynamics of contacts between pairs of individuals. We aimed to analyse contact networks with respect to sociodemographic and geographical characteristics, vaccine-induced immunity and self-reported symptoms.

Methods

In 2014, volunteers from two large participatory surveillance panels in the Netherlands and Belgium were invited for a survey. Participants were asked to record numbers of contacts at different locations and self-reported influenza-like-illness symptoms, and to invite 4 individuals they had met face to face in the preceding 2 weeks. We calculated correlations between linked individuals to investigate mixing patterns.

Results

In total 1560 individuals completed the survey who reported in total 30591 contact persons; 488 recruiter-recruit pairs were analysed. Recruitment was assortative by age, education, household size, influenza vaccination status and sentiments, indicating that participants tended to recruit contact persons similar to themselves. We also found assortative recruitment by symptoms, reaffirming our objective of sampling contact persons whom a participant may infect or by whom a participant may get infected in case of an outbreak. Recruitment was random by sex and numbers of contact persons. Relationships between pairs were influenced by the spatial distribution of peer recruitment.

Conclusions

Although complex mechanisms influence online peer recruitment, the observed statistical relationships reflected the observed contact network patterns in the general population relevant for the transmission of respiratory pathogens. This provides useful and innovative input for predictive epidemic models relying on network information.

Electronic supplementary material

The online version of this article (doi:10.1186/s12879-015-1250-z) contains supplementary material, which is available to authorized users.