Incidence and risk factors for influenza-like-illness in the UK: online surveillance using Flusurvey.

BACKGROUND: Influenza and Influenza-like-illness (ILI) represents a substantial public health problem, but it is difficult to measure the overall burden as many cases do not access health care. Community cohorts have the advantage of not requiring individuals to present at hospitals and surgeries and therefore can potentially monitor a wider variety of cases. This study reports on the incidence and risk factors for ILI in the UK as measured using Flusurvey, an internet-based open community cohort. METHODS: Upon initial online registration participants were asked background characteristics, and every week were asked to complete a symptoms survey. We compared the representativeness of our sample to the overall population. We used two case definitions of ILI, which differed in whether fever/chills was essential. We calculated ILI incidence week by week throughout the season, and investigated risk factors associated with ever reporting ILI over the course of the season. Risk factor analysis was conducted using binomial regression. RESULTS: 5943 participants joined the survey, and 4532 completed the symptoms survey at least twice. Participants who filled in symptoms surveys at least twice filled in a median of nine symptoms surveys over the course of the study. 46.1% of participants reported at least one episode of ILI, and 6.0% of all reports were positive for ILI. Females had slightly higher incidence, and individuals over 65 had the lowest incidence. Incidence peaked just before Christmas and declined dramatically during school holidays. Multivariate regression showed that, for both definitions of ILI considered, being female, unvaccinated, having underlying health issues, having contact with children, being aged between 35 and 64, and being a smoker were associated with the highest risk of reporting an ILI. The use of public transport was not associated with an increased risk of ILI. CONCLUSIONS: Our results show that internet based surveillance can be used to measure ILI and understand risk factors. Vaccination is shown to be linked to a reduced risk of reporting ILI. Taking public transport does not increase the risk of reporting ILI. Flusurvey and other participatory surveillance techniques can be used to provide reliable information to policy makers in nearly real-time.

The impact of illness on social networks: implications for transmission and control of influenza.

We expect social networks to change as a result of illness, but social contact data are generally collected from healthy persons. Here we quantified the impact of influenza-like illness on social mixing patterns. We analyzed the contact patterns of persons from England measured when they were symptomatic with influenza-like illness during the 2009 A/H1N1pdm influenza epidemic (2009-2010) and again 2 weeks later when they had recovered. Illness was associated with a reduction in the number of social contacts, particularly in settings outside the home, reducing the reproduction number to about one-quarter of the value it would otherwise have taken. We also observed a change in the age distribution of contacts. By comparing the expected age distribution of cases resulting from transmission by (a)symptomatic persons with incidence data, we estimated the contribution of both groups to transmission. Using this, we calculated the fraction of transmission resulting from (a)symptomatic persons, assuming equal duration of infectiousness. We estimated that 66% of transmission was attributable to persons with symptomatic disease (95% confidence interval: 0.23, 1.00). This has important implications for control: Treating symptomatic persons with antiviral agents or encouraging home isolation would be expected to have a major impact on transmission, particularly since the reproduction number for this strain was low.

Measured dynamic social contact patterns explain the spread of H1N1v influenza.

Patterns of social mixing are key determinants of epidemic spread. Here we present the results of an internet-based social contact survey completed by a cohort of participants over 9,000 times between July 2009 and March 2010, during the 2009 H1N1v influenza epidemic. We quantify the changes in social contact patterns over time, finding that school children make 40% fewer contacts during holiday periods than during term time. We use these dynamically varying contact patterns to parameterise an age-structured model of influenza spread, capturing well the observed patterns of incidence; the changing contact patterns resulted in a fall of approximately 35% in the reproduction number of influenza during the holidays. This work illustrates the importance of including changing mixing patterns in epidemic models. We conclude that changes in contact patterns explain changes in disease incidence, and that the timing of school terms drove the 2009 H1N1v epidemic in the UK. Changes in social mixing patterns can be usefully measured through simple internet-based surveys.

A class of pairwise models for epidemic dynamics on weighted networks.

In this paper, we study the SIS (susceptible-infected-susceptible) and SIR (susceptible-infected-removed) epidemic models on undirected, weighted networks by deriving pairwise-type approximate models coupled with individual-based network simulation. Two different types of theoretical/synthetic weighted network models are considered. Both start from non-weighted networks with fixed topology followed by the allocation of link weights in either (i) random or (ii) fixed/deterministic way. The pairwise models are formulated for a general discrete distribution of weights, and these models are then used in conjunction with stochastic network simulations to evaluate the impact of different weight distributions on epidemic thresholds and dynamics in general. For the SIR model, the basic reproductive ratio R0 is computed, and we show that (i) for both network models R0 is maximised if all weights are equal, and (ii) when the two models are 'equally-matched', the networks with a random weight distribution give rise to a higher R0 value. The models with different weight distributions are also used to explore the agreement between the pairwise and simulation models for different parameter combinations.

Coughs, colds and "freshers' flu" survey in the University of Cambridge, 2007-2008.

Universities provide many opportunities for the spread of infectious respiratory illnesses. Students are brought together into close proximity from all across the world and interact with one another in their accommodation, through lectures and small group teaching and in social settings. The COVID-19 global pandemic has highlighted the need for sufficient data to help determine which of these factors are important for infectious disease transmission in universities and hence control university morbidity as well as community spillover. We describe the data from a previously unpublished self-reported university survey of coughs, colds and influenza-like symptoms collected in Cambridge, UK, during winter 2007-2008. The online survey collected information on symptoms and socio-demographic, academic and lifestyle factors. There were 1076 responses, 97% from University of Cambridge students (5.7% of the total university student population), 3% from staff and <1% from other participants, reporting onset of symptoms between September 2007 and March 2008. Undergraduates are seen to report symptoms earlier in the term than postgraduates; differences in reported date of symptoms are also seen between subjects and accommodation types, although these descriptive results could be confounded by survey biases. Despite the historical and exploratory nature of the study, this is one of few recent detailed datasets of influenza-like infection in a university context and is especially valuable to share now to improve understanding of potential transmission dynamics in universities during the current COVID-19 pandemic.

Using an online survey of healthcare-seeking behaviour to estimate the magnitude and severity of the 2009 H1N1v influenza epidemic in England.

BACKGROUND: During the 2009 H1N1v influenza epidemic, the total number of symptomatic cases was estimated by combining influenza-like illness (ILI) consultations, virological surveillance and assumptions about healthcare-seeking behaviour. Changes in healthcare-seeking behaviour due to changing scientific information, media coverage and public anxiety, were not included in case estimates. The purpose of the study was to improve estimates of the number of symptomatic H1N1v cases and the case fatality rate (CFR) in England by quantifying healthcare-seeking behaviour using an internet-based survey carried out during the course of the 2009 H1N1v influenza epidemic. METHODS: We used an online survey that ran continuously from July 2009 to March 2010 to estimate the proportion of ILI cases that sought healthcare during the 2009 H1N1v influenza epidemic. We used dynamic age- and gender-dependent measures of healthcare-seeking behaviour to re-interpret consultation numbers and estimate the true number of cases of symptomatic ILI in 2009 and the case fatality rate (CFR). RESULTS: There were significant differences between age groups in healthcare usage. From the start to the end of the epidemic, the percentage of individuals with influenza-like symptoms who sought medical attention decreased from 43% to 32% (p < 0.0001). Adjusting official numbers accordingly, we estimate that there were 1.1 million symptomatic cases in England, over 320,000 (40%) more cases than previously estimated and that the autumn epidemic wave was 45% bigger than previously thought. Combining symptomatic case numbers with reported deaths leads to a reduced overall CFR estimate of 17 deaths per 100,000 cases, with the largest reduction in adults. CONCLUSIONS: Active surveillance of healthcare-seeking behaviour, which can be achieved using novel data collection methods, is vital for providing accurate real-time estimates of epidemic size and disease severity. The differences in healthcare-seeking between different population groups and changes over time have significant implications for estimates of total case numbers and the case fatality rate.

Assessing the role of contact tracing in a suspected H7N2 influenza A outbreak in humans in Wales.

BACKGROUND: The detailed analysis of an outbreak database has been undertaken to examine the role of contact tracing in controlling an outbreak of possible avian influenza in humans. The outbreak, initiating from the purchase of infected domestic poultry, occurred in North Wales during May and June 2007. During this outbreak, extensive contact tracing was carried out. Following contact tracing, cases and contacts believed to be at risk of infection were given treatment/prophylaxis. METHODS: We analyse the database of cases and their contacts identified for the purposes of contact tracing in relation to both the contact tracing burden and effectiveness. We investigate the distribution of numbers of contacts identified, and use network structure to explore the speed with which treatment/prophylaxis was made available and to estimate the risk of transmission in different settings. RESULTS: Fourteen cases of suspected H7N2 influenza A in humans were associated with a confirmed outbreak among poultry in May-June 2007. The contact tracing dataset consisted of 254 individuals (cases and contacts, of both poultry and humans) who were linked through a network of social contacts. Of these, 102 individuals were given treatment or prophylaxis. Considerable differences between individuals' contact patterns were observed. Home and workplace encounters were more likely to result in transmission than encounters in other settings. After an initial delay, while the outbreak proceeded undetected, contact tracing rapidly caught up with the cases and was effective in reducing the time between onset of symptoms and treatment/prophylaxis. CONCLUSIONS: Contact tracing was used to link together the individuals involved in this outbreak in a social network, allowing the identification of the most likely paths of transmission and the risks of different types of interactions to be assessed. The outbreak highlights the substantial time and cost involved in contact tracing, even for an outbreak affecting few individuals. However, when sufficient resources are available, contact tracing enables cases to be identified before they result in further transmission and thus possibly assists in preventing an outbreak of a novel virus.

Internet-based surveillance of Influenza-like-illness in the UK during the 2009 H1N1 influenza pandemic.

BACKGROUND: Internet-based surveillance systems to monitor influenza-like illness (ILI) have advantages over traditional (physician-based) reporting systems, as they can potentially monitor a wider range of cases (i.e. including those that do not seek care). However, the requirement for participants to have internet access and to actively participate calls into question the representativeness of the data. Such systems have been in place in a number of European countries over the last few years, and in July 2009 this was extended to the UK. Here we present results of this survey with the aim of assessing the reliability of the data, and to evaluate methods to correct for possible biases. METHODS: Internet-based monitoring of ILI was launched near the peak of the first wave of the UK H1N1v influenza pandemic. We compared the recorded ILI incidence with physician-recorded incidence and an estimate of the true number of cases over the course of the epidemic. We also compared overall attack rates. The effect of using different ILI definitions and alternative denominator assumptions on incidence estimates was explored. RESULTS: The crude incidence measured by the internet-based system appears to be influenced by individuals who participated only once in the survey and who appeared more likely to be ill. This distorted the overall incidence trend. Concentrating on individuals who reported more than once results in a time series of ILI incidence that matches the trend of case estimates reasonably closely, with a correlation of 0.713 (P-value: 0.0001, 95% CI: 0.435, 0.867). Indeed, the internet-based system appears to give a better estimate of the relative height of the two waves of the UK pandemic than the physician-recorded incidence. The overall attack rate is, however, higher than other estimates, at about 16% when compared with a model-based estimate of 6%. CONCLUSION: Internet-based monitoring of ILI can capture the trends in case numbers if appropriate weighting is used to correct for differential response. The overall level of incidence is, however, difficult to measure. Internet-based systems may be a useful adjunct to existing ILI surveillance systems as they capture cases that do not necessarily contact health care. However, further research is required before they can be used to accurately assess the absolute level of incidence in the community.

Dynamic social networks and the implications for the spread of infectious disease.

Understanding the nature of human contact patterns is crucial for predicting the impact of future pandemics and devising effective control measures. However, few studies provide a quantitative description of the aspects of social interactions that are most relevant to disease transmission. Here, we present the results from a detailed diary-based survey of casual (conversational) and close contact (physical) encounters made by a small peer group of 49 adults who recorded 8,661 encounters with 3,528 different individuals over 14 non-consecutive days. We find that the stability of interactions depends on the intimacy of contact and social context. Casual contact encounters mostly occur in the workplace and are predominantly irregular, while close contact encounters mostly occur at home or in social situations and tend to be more stable. Simulated epidemics of casual contact transmission involve a large number of non-repeated encounters, and the social network is well captured by a random mixing model. However, the stability of the social network should be taken into account for close contact infections. Our findings have implications for the modelling of human epidemics and planning pandemic control policies based on social distancing methods.

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.

Coexistence and specialization of pathogen strains on contact networks.

The coexistence of different pathogen strains has implications for pathogen variability and disease control and has been explained in a number of different ways. We use contact networks, which represent interactions between individuals through which infection could be transmitted, to investigate strain coexistence. For sexually transmitted diseases the structure of contact networks has received detailed study and has been shown to be a vital determinant of the epidemiological dynamics. By using analytical pairwise models and stochastic simulations, we demonstrate that network structure also has a profound influence on the interaction between pathogen strains. In particular, when the population is serially monogamous, fully cross-reactive strains can coexist, with different strains dominating in network regions with different characteristics. Furthermore, we observe specialization of different strains in different risk groups within the network, suggesting the existence of diverging evolutionary pressures.

Networks and epidemic models.

Networks and the epidemiology of directly transmitted infectious diseases are fundamentally linked. The foundations of epidemiology and early epidemiological models were based on population wide random-mixing, but in practice each individual has a finite set of contacts to whom they can pass infection; the ensemble of all such contacts forms a 'mixing network'. Knowledge of the structure of the network allows models to compute the epidemic dynamics at the population scale from the individual-level behaviour of infections. Therefore, characteristics of mixing networks-and how these deviate from the random-mixing norm-have become important applied concerns that may enhance the understanding and prediction of epidemic patterns and intervention measures. Here, we review the basis of epidemiological theory (based on random-mixing models) and network theory (based on work from the social sciences and graph theory). We then describe a variety of methods that allow the mixing network, or an approximation to the network, to be ascertained. It is often the case that time and resources limit our ability to accurately find all connections within a network, and hence a generic understanding of the relationship between network structure and disease dynamics is needed. Therefore, we review some of the variety of idealized network types and approximation techniques that have been utilized to elucidate this link. Finally, we look to the future to suggest how the two fields of network theory and epidemiological modelling can deliver an improved understanding of disease dynamics and better public health through effective disease control.

School's Out: Seasonal Variation in the Movement Patterns of School Children.

School children are core groups in the transmission of many common infectious diseases, and are likely to play a key role in the spatial dispersal of disease across multiple scales. However, there is currently little detailed information about the spatial movements of this epidemiologically important age group. To address this knowledge gap, we collaborated with eight secondary schools to conduct a survey of movement patterns of school pupils in primary and secondary schools in the United Kingdom. We found evidence of a significant change in behaviour between term time and holidays, with term time weekdays characterised by predominately local movements, and holidays seeing much broader variation in travel patterns. Studies that use mathematical models to examine epidemic transmission and control often use adult commuting data as a proxy for population movements. We show that while these data share some features with the movement patterns reported by school children, there are some crucial differences between the movements of children and adult commuters during both term-time and holidays.

Nine challenges in incorporating the dynamics of behaviour in infectious diseases models.

Traditionally, the spread of infectious diseases in human populations has been modelled with static parameters. These parameters, however, can change when individuals change their behaviour. If these changes are themselves influenced by the disease dynamics, there is scope for mechanistic models of behaviour to improve our understanding of this interaction. Here, we present challenges in modelling changes in behaviour relating to disease dynamics, specifically: how to incorporate behavioural changes in models of infectious disease dynamics, how to inform measurement of relevant behaviour to parameterise such models, and how to determine the impact of behavioural changes on observed disease dynamics.

Modeling infectious disease dynamics in the complex landscape of global health.

Despite some notable successes in the control of infectious diseases, transmissible pathogens still pose an enormous threat to human and animal health. The ecological and evolutionary dynamics of infections play out on a wide range of interconnected temporal, organizational, and spatial scales, which span hours to months, cells to ecosystems, and local to global spread. Moreover, some pathogens are directly transmitted between individuals of a single species, whereas others circulate among multiple hosts, need arthropod vectors, or can survive in environmental reservoirs. Many factors, including increasing antimicrobial resistance, increased human connectivity and changeable human behavior, elevate prevention and control from matters of national policy to international challenge. In the face of this complexity, mathematical models offer valuable tools for synthesizing information to understand epidemiological patterns, and for developing quantitative evidence for decision-making in global health.

Contact tracing and disease control.

Contact tracing, followed by treatment or isolation, is a key control measure in the battle against infectious diseases. It is an extreme form of locally targeted control, and as such has the potential to be highly efficient when dealing with low numbers of cases. For this reason it is frequently used to combat sexually transmitted diseases and new invading pathogens. Accurate modelling of contact tracing requires explicit information about the disease-transmission pathways from each individual, and hence the network of contacts. Here, pairwise-approximation methods and full stochastic simulations are used to investigate the utility of contact tracing. A simple relationship is found between the efficiency of contact tracing necessary for eradication and the basic reproductive ratio of the disease. This holds for a wide variety of realistic situations including heterogeneous networks containing core-groups or super-spreaders, and asymptomatic individuals. Clustering (transitivity) within the transmission network is found to destroy the relationship, requiring lower efficiency than predicted.

Monogamous networks and the spread of sexually transmitted diseases.

Patterns of sexual mixing and heterogeneity in the number of sexual partners can have a huge effect on the spread of a sexually transmitted disease (STD). The sexual mixing network identifies all partnerships within a population over a given period and is a powerful tool in the study of such infections. Previous models assumed all links within the network to be concurrent active partnerships. We present a novel modelling approach in which we adapt the notion of a sexual contact network to a monogamous population by allowing the nature of the links to change. We use the underlying network to represent potential sexual partnerships, only some of which are active at any one time. Thus serial monogamy can be modelled while maintaining the patterns of mixing displayed by the population.

Integrating contact network structure into tuberculosis epidemiology in meerkats in South Africa: Implications for control.

Empirical studies that integrate information on host contact patterns with infectious disease transmission over time are rare. The aims of this study were to determine the relative importance of intra-group social interactions in the transmission of tuberculosis (TB; Mycobacterium bovis infection) in a population of wild meerkats (Suricata suricatta) in South Africa, and to use this information to propose an evidence-based intervention strategy to manage this disease. Detailed behavioural observations of all members of eight meerkat groups (n=134 individuals) were made over 24 months from January 2006 to December 2007. Social network analysis of three types of interaction (aggression, foraging competitions and grooming) revealed social structure to be very stable over time. Clustering of interactions was positively correlated with group size for both aggression (r=0.73) and grooming interactions (r=0.71), suggesting that infections may spread locally within clusters of interacting individuals but be limited from infecting all members of large groups by an apparent threshold in connections between different clusters. Repeated biological sampling every three months of all members of one social group (n=37 meerkats) was undertaken to quantify individual changes in M. bovis infection status. These empirical data were used to construct a dynamic network model of TB transmission within a meerkat group. The results indicated that grooming (both giving and receiving) was more likely than aggression to be correlated with M. bovis transmission and that groomers were at higher risk of infection than groomees. Intervention strategies for managing TB in meerkats that focus on those individuals engaging in the highest amount of grooming are therefore proposed.

The impact of school holidays on the social mixing patterns of school children.

School holidays are recognised to be of great epidemiological importance for a wide range of infectious diseases; this is postulated to be because the social mixing patterns of school children - a key population group - change significantly during the holiday period. However, there is little direct quantitative evidence to confirm this belief. Here, we present the results of a prospective survey designed to provide a detailed comparison of social mixing patterns of school children during school terms and during the school holidays. Paired data were collected, with participants recording their social contacts once during term time and once during the holiday period. We found that the daily number of recorded encounters approximately halved during the holidays, and that the number of close contact encounters fell by approximately one third. The holiday period also saw a change in the age structure of children's social contacts, with far fewer contacts of their own age, but an increase in the number of encounters with adults, particularly older adults. A greater amount of mixing between children at different schools was recorded during the holiday. We suggest, therefore, that whilst infections may spread rapidly within schools during term time, in the holiday period there are increased opportunities for transmission to other schools and other age groups.

Networks of influence and infection: parental choices and childhood disease.

Clusters of unvaccinated individuals are at risk of outbreaks of infection. When an individual's decision to choose vaccination is influenced by the choices of his social group, such clusters can readily arise. However, when the interactions that influence decision-making and those that permit the transmission of infection are different--for instance, when parents make vaccination decisions on behalf of their children--it is unclear how large the impact of this social influence will be. Here we use a modelling approach to represent social influence within a network of parents and the transmission of infection through a network of children. We show that the effect of social influence depends on the amount of overlap between the two different networks; large overlap means that clusters of parents who choose not to vaccinate are likely to have interacting children, generating clusters of unvaccinated children. Spatially local connections can further increase the impact of social influence. Outbreaks are most likely when parents who do not vaccinate have children who interact.