Modelling seasonal influenza: the role of weather and punctuated antigenic drift

Forecasting seasonal influenza activity in Canada-Comparing seasonal Auto-Regressive integrated moving average and artificial neural network approaches for public health preparedness

INTRODUCTION

Public health preparedness is based on timely and accurate information. Time series forecasting using disease surveillance data is an important aspect of preparedness. This study compared two approaches of time series forecasting: seasonal auto-regressive integrated moving average (SARIMA) modelling and the artificial neural network (ANN) algorithm. The goal was to model weekly seasonal influenza activity in Canada using SARIMA and compares its predictive accuracy, based on root mean square prediction error (RMSE) and mean absolute prediction error (MAE), to that of an ANN.

METHODS

An initial SARIMA model was fit using automated model selection by minimizing the Akaike information criterion (AIC). Further inspection of the autocorrelation function and partial autocorrelation function led to 'manual' model improvements. ANNs were trained iteratively, using an automated process to minimize the RMSE and MAE.

RESULTS

A total of 378, 462 cases of influenza was reported in Canada from the 2010-2011 influenza season to the end of the 2019-2020 influenza season, with an average yearly incidence risk of 20.02 per 100,000 population. Automated SARIMA modelling was the better method in terms of forecasting accuracy (per RMSE and MAE). However, the ANN correctly predicted the peak week of disease incidence while the other models did not.

CONCLUSION

Both the ANN and SARIMA models have shown to be capable tools in forecasting seasonal influenza activity in Canada. It was shown that applying both in tandem is beneficial, SARIMA better forecasted overall incidence while ANN correctly predicted the peak week.

The effects of weather and mobility on respiratory viruses dynamics before and during the COVID-19 pandemic in the USA and Canada

The flu season is caused by a combination of different pathogens, including influenza viruses (IVS), that cause the flu, and non-influenza respiratory viruses (NIRVs), that cause common colds or influenza-like illness. These viruses exhibit similar dynamics and meteorological conditions have historically been regarded as a principal modulator of their epidemiology, with outbreaks in the winter and almost no circulation during the summer, in temperate regions. However, after the emergence of SARS-CoV2, in late 2019, the dynamics of these respiratory viruses were strongly perturbed worldwide: some infections displayed near-eradication, while others experienced temporal shifts or occurred "off-season". This disruption raised questions regarding the dominant role of weather while also providing an unique opportunity to investigate the roles of different determinants on the epidemiological dynamics of IVs and NIRVs. Here, we employ statistical analysis and modelling to test the effects of weather and mobility in viral dynamics, before and during the COVID-19 pandemic. Leveraging epidemiological surveillance data on several respiratory viruses, from Canada and the USA, from 2016 to 2023, we found that whereas in the pre-COVID-19 pandemic period, weather had a strong effect, in the pandemic period the effect of weather was strongly reduced and mobility played a more relevant role. These results, together with previous studies, indicate that behavioral changes resulting from the non-pharmacological interventions implemented to control SARS-CoV2, interfered with the dynamics of other respiratory viruses, and that the past dynamical equilibrium was disturbed, and perhaps permanently altered, by the COVID-19 pandemic.

A discrete-time susceptible-infectious-recovered-susceptible model for the analysis of influenza data

We develop a discrete time compartmental model to describe the spread of seasonal influenza virus. As time and disease state variables are assumed to be discrete, this model is considered to be a discrete time, stochastic, Susceptible-Infectious-Recovered-Susceptible (DT-SIRS) model, where weekly counts of disease are assumed to follow a Poisson distribution. We allow the disease transmission rate to also vary over time, and the disease can only be reintroduced after extinction if there is a contact with infected individuals from other host populations. To capture the variability of influenza activities from one season to the next, we define the seasonality with a 4-week period effect that may change over years. We examine three different transmission rates and compare their performance to that of existing approaches. Even though there is limited information for susceptible and recovered individuals, we demonstrate that the simple models for transmission rates effectively capture the behaviour of the disease dynamics. We use a Bayesian approach for inference. The framework is applied in an analysis of the temporal spread of influenza in the province of Manitoba, Canada, 2012-2015.

Competitive evolution of H1N1 and H3N2 influenza viruses in the United States: A mathematical modeling study

Seasonal influenza causes vast public health and economic impact globally. The prevention and control of the annual epidemics remain a challenge due to the antigenic evolution of the viruses. Here, we presented a novel modeling framework based on changes in amino acid sequences and relevant epidemiological data to retrospectively investigate the competitive evolution and transmission of H1N1 and H3N2 influenza viruses in the United States during October 2002 and April 2019. To do so, we estimated the time-varying disease transmission rate from the reported influenza cases and the time-varying antigenic change rate of the viruses from the changes in amino acid sequences. By incorporating the time-varying antigenic change rate into the transmission models, we found that the models could capture the evolutionary transmission dynamics of influenza viruses in the United States. Our modeling results also showed that the antigenic change of the virus plays an essential role in seasonal influenza dynamics.

Mechanisms for the circulation of influenza A(H3N2) in China: A spatiotemporal modelling study

Circulation of seasonal influenza is the product of complex interplay among multiple drivers, yet characterizing the underlying mechanism remains challenging. Leveraging the diverse seasonality of A(H3N2) virus and abundant climatic space across regions in China, we quantitatively investigated the relative importance of population susceptibility, climatic factors, and antigenic change on the dynamics of influenza A(H3N2) through an integrative modelling framework. Specifically, an absolute humidity driven multiscale transmission model was constructed for the 2013/2014, 2014/2015 and 2016/2017 influenza seasons that were dominated by influenza A(H3N2). We revealed the variable impact of absolute humidity on influenza transmission and differences in the occurring timing and magnitude of antigenic change for those three seasons. Overall, the initial population susceptibility, climatic factors, and antigenic change explained nearly 55% of variations in the dynamics of influenza A(H3N2). Specifically, the additional variation explained by the initial population susceptibility, climatic factors, and antigenic change were at 33%, 26%, and 48%, respectively. The vaccination program alone failed to fully eliminate the summer epidemics of influenza A(H3N2) and non-pharmacological interventions were needed to suppress the summer circulation. The quantitative understanding of the interplay among driving factors on the circulation of influenza A(H3N2) highlights the importance of simultaneous monitoring of fluctuations for related factors, which is crucial for precise and targeted prevention and control of seasonal influenza.

Antigenic drift and epidemiological severity of seasonal influenza in Canada

Seasonal influenza epidemics circulate globally every year with varying levels of severity. One of the major drivers of this seasonal variation is thought to be the antigenic drift of influenza viruses, resulting from the accumulation of mutations in viral surface proteins. In this study, we aimed to investigate the association between the genetic drift of seasonal influenza viruses (A/H1N1, A/H3N2 and B) and the epidemiological severity of seasonal epidemics within a Canadian context. We obtained hemagglutinin protein sequences collected in Canada between the 2006/2007 and 2019/2020 flu seasons from GISAID and calculated Hamming distances in a sequence-based approach to estimating inter-seasonal antigenic differences. We also gathered epidemiological data on cases, hospitalizations and deaths from national surveillance systems and other official sources, as well as vaccine effectiveness estimates to address potential effect modification. These aggregate measures of disease severity were integrated into a single seasonal severity index. We performed linear regressions of our severity index with respect to the inter-seasonal antigenic distances, controlling for vaccine effectiveness. We did not find any evidence of a statistical relationship between antigenic distance and seasonal influenza severity in Canada. Future studies may need to account for additional factors, such as co-circulation of other respiratory pathogens, population imprinting, cohort effects and environmental parameters, which may drive seasonal influenza severity.

Effects of climatic factors on the prevalence of influenza virus infection in Cheonan, Korea

Big data can be used to correlate diseases and climatic factors. The prevalence of influenza (flu) virus, accounting for a large proportion of respiratory infections, suggests that the effect of climate variables according to seasonal dynamics of influenza virus infections should be investigated. Here, trends in flu virus detection were analyzed using data from 9,010 tests performed between January 2012 and December 2018 at Dankook University Hospital, Cheonan, Korea. We compared the detection of the flu virus in Cheonan area and its association with climate change. The flu virus detection rate was 9.9% (894/9,010), and the detection rate was higher for flu virus A (FLUAV; 6.9%) than for flu virus B (FLUBV; 3.0%). Both FLUAV and FLUBV infections are considered an epidemic each year. We identified 43.1% (n = 385) and 35.0% (n = 313) infections in children aged < 10 years and adults aged > 60 years, respectively. The combination of these age groups encompassed 78.1% (n = 698/894) of the total data. Flu virus infections correlated with air temperature, relative humidity, vapor pressure, atmospheric pressure, particulate matter, and wind chill temperature (P < 0.001). However, the daily temperature range did not significantly correlate with the flu detection results. This is the first study to identify the relationship between long-term flu virus infection with temperature in the temperate region of Cheonan.

Mutation induced infection waves in diseases like COVID-19

After more than 6 million deaths worldwide, the ongoing vaccination to conquer the COVID-19 disease is now competing with the emergence of increasingly contagious mutations, repeatedly supplanting earlier strains. Following the near-absence of historical examples of the long-time evolution of infectious diseases under similar circumstances, models are crucial to exemplify possible scenarios. Accordingly, in the present work we systematically generalize the popular susceptible-infected-recovered model to account for mutations leading to repeatedly occurring new strains, which we coarse grain based on tools from statistical mechanics to derive a model predicting the most likely outcomes. The model predicts that mutations can induce a super-exponential growth of infection numbers at early times, which self-amplify to giant infection waves which are caused by a positive feedback loop between infection numbers and mutations and lead to a simultaneous infection of the majority of the population. At later stages-if vaccination progresses too slowly-mutations can interrupt an ongoing decrease of infection numbers and can cause infection revivals which occur as single waves or even as whole wave trains featuring alternative periods of decreasing and increasing infection numbers. This panorama of possible mutation-induced scenarios should be tested in more detailed models to explore their concrete significance for specific infectious diseases. Further, our results might be useful for discussions regarding the importance of a release of vaccine-patents to reduce the risk of mutation-induced infection revivals but also to coordinate the release of measures following a downwards trend of infection numbers.

Mitigating bias in estimating epidemic severity due to heterogeneity of epidemic onset and data aggregation

Outbreaks of infectious diseases, such as influenza, are a major societal burden. Mitigation policies during an outbreak or pandemic are guided by the analysis of data of ongoing or preceding epidemics. The reproduction number, R0, defined as the expected number of secondary infections arising from a single individual in a population of susceptibles is critical to epidemiology. For typical compartmental models such as the Susceptible-Infected-Recovered (SIR) R0 represents the severity of an epidemic. It is an estimate of the early-stage growth rate of an epidemic and is an important threshold parameter used to gain insights into the spread or decay of an outbreak. Models typically use incidence counts as indicators of cases within a single large population; however, epidemic data are the result of a hierarchical aggregation, where incidence counts from spatially separated monitoring sites (or sub-regions) are pooled and used to infer R0. Is this aggregation approach valid when the epidemic has different dynamics across the regions monitored? We characterize bias in the estimation of R0 from a merged data set when the epidemics of the sub-regions, used in the merger, exhibit delays in onset. We propose a method to mitigate this bias, and study its efficacy on synthetic data as well as real-world influenza and COVID-19 data.

Impact of biometeorological conditions and air pollution on influenza-like illnesses incidence in Warsaw

In order to assess the influence of atmospheric conditions and particulate matter (PM) on the seasonally varying incidence of influenza-like illnesses (ILI) in the capital of Poland-Warsaw, we analysed time series of ILI reported for the about 1.75 million residents in total and for different age groups in 288 approximately weekly periods, covering 6 years 2013-2018. Using Poisson regression, we predicted ILI by the Universal Thermal Climate Index (UTCI) as biometeorological indicator, and by PM2.5 and PM10, respectively, as air quality measures accounting for lagged effects spanning up to 3 weeks. Excess ILI incidence after adjusting for seasonal and annual trends was calculated by fitting generalized additive models. ILI morbidity increased with rising PM concentrations, for both PM2.5 and PM10, and with cooler atmospheric conditions as indicated by decreasing UTCI. While the PM effect focused on the actual reporting period, the atmospheric influence exhibited a more evenly distributed lagged effect pattern over the considered 3-week period. Though ILI incidence adjusted for population size significantly declined with age, age did not significantly modify the effect sizes of both PM and UTCI. These findings contribute to better understanding environmental conditionings of influenza seasonality in a temperate climate. This will be beneficial to forecasting future dynamics of ILI and to planning clinical and public health resources under climate change scenarios.

Respiratory Co-Infections: Modulators of SARS-CoV-2 Patients' Clinical Sub-Phenotype

Co-infection with ancillary pathogens is a significant modulator of morbidity and mortality in infectious diseases. There have been limited reports of co-infections accompanying SARS-CoV-2 infections, albeit lacking India specific study. The present study has made an effort toward elucidating the prevalence, diversity and characterization of co-infecting respiratory pathogens in the nasopharyngeal tract of SARS-CoV-2 positive patients. Two complementary metagenomics based sequencing approaches, Respiratory Virus Oligo Panel (RVOP) and Holo-seq, were utilized for unbiased detection of co-infecting viruses and bacteria. The limited SARS-CoV-2 clade diversity along with differential clinical phenotype seems to be partially explained by the observed spectrum of co-infections. We found a total of 43 bacteria and 29 viruses amongst the patients, with 18 viruses commonly captured by both the approaches. In addition to SARS-CoV-2, Human Mastadenovirus, known to cause respiratory distress, was present in a majority of the samples. We also found significant differences of bacterial reads based on clinical phenotype. Of all the bacterial species identified, ∼60% have been known to be involved in respiratory distress. Among the co-pathogens present in our sample cohort, anaerobic bacteria accounted for a preponderance of bacterial diversity with possible role in respiratory distress. Clostridium botulinum, Bacillus cereus and Halomonas sp. are anaerobes found abundantly across the samples. Our findings highlight the significance of metagenomics based diagnosis and detection of SARS-CoV-2 and other respiratory co-infections in the current pandemic to enable efficient treatment administration and better clinical management. To our knowledge this is the first study from India with a focus on the role of co-infections in SARS-CoV-2 clinical sub-phenotype.

The relationship between cyclonic weather regimes and seasonal influenza over the Eastern Mediterranean

The prediction of the occurrence of infectious diseases is of crucial importance for public health, as clearly seen in the ongoing COVID-19 pandemic. Here, we analyze the relationship between the occurrence of a winter low-pressure weather regime - Cyprus Lows - and the seasonal Influenza in the Eastern Mediterranean. We find that the weekly occurrence of Cyprus Lows is significantly correlated with clinical seasonal Influenza in Israel in recent years (R = 0.91; p < .05). This result remains robust when considering a complementary analysis based on Google Trends data for Israel, the Palestinian Authority and Jordan. The weekly occurrence of Cyprus Lows precedes the onset and maximum of Influenza occurrence by about one to two weeks (R = 0.88; p < .05 for the maximum occurrence), and closely follows their timing in eight out of ten years (2008-2017). Since weather regimes such as Cyprus Lows are more robustly predicted in weather and climate models than individual climate variables, we conclude that the weather regime approach can be used to develop tools for estimating the compatibility of the transmission environment for Influenza occurrence in a warming world. Furthermore, this approach may be applied to other regions and climate sensitive diseases. This study is a new cross-border inter-disciplinary regional collaboration for appropriate adaptation to climate change in the Eastern Mediterranean.

The local stability of a modified multi-strain SIR model for emerging viral strains

We study a novel multi-strain SIR epidemic model with selective immunity by vaccination. A newer strain is made to emerge in the population when a preexisting strain has reached equilbrium. We assume that this newer strain does not exhibit cross-immunity with the original strain, hence those who are vaccinated and recovered from the original strain become susceptible to the newer strain. Recent events involving the COVID-19 virus shows that it is possible for a viral strain to emerge from a population at a time when the influenza virus, a well-known virus with a vaccine readily available, is active in a population. We solved for four different equilibrium points and investigated the conditions for existence and local stability. The reproduction number was also determined for the epidemiological model and found to be consistent with the local stability condition for the disease-free equilibrium.

Dynamic interactions of influenza viruses in Hong Kong during 1998-2018

Influenza epidemics cause substantial morbidity and mortality every year worldwide. Currently, two influenza A subtypes, A(H1N1) and A(H3N2), and type B viruses co-circulate in humans and infection with one type/subtype could provide cross-protection against the others. However, it remains unclear how such ecologic competition via cross-immunity and antigenic mutations that allow immune escape impact influenza epidemic dynamics at the population level. Here we develop a comprehensive model-inference system and apply it to study the evolutionary and epidemiological dynamics of the three influenza types/subtypes in Hong Kong, a city of global public health significance for influenza epidemic and pandemic control. Utilizing long-term influenza surveillance data since 1998, we are able to estimate the strength of cross-immunity between each virus-pairs, the timing and frequency of punctuated changes in population immunity in response to antigenic mutations in influenza viruses, and key epidemiological parameters over the last 20 years including the 2009 pandemic. We find evidence of cross-immunity in all types/subtypes, with strongest cross-immunity from A(H1N1) against A(H3N2). Our results also suggest that A(H3N2) may undergo antigenic mutations in both summers and winters and thus monitoring the virus in both seasons may be important for vaccine development. Overall, our study reveals intricate epidemiological interactions and underscores the importance of simultaneous monitoring of population immunity, incidence rates, and viral genetic and antigenic changes.

Analyzing influenza outbreaks in Russia using an age-structured dynamic transmission model

In this study, we addressed the ability of a minimalistic SEIR model to satisfactorily describe influenza outbreak dynamics in Russian settings. For that purpose, we calibrated an age-specific influenza dynamics model to Russian acute respiratory infection (ARI) incidence data over 2009-2016 and assessed the variability of proportion of non-immune individuals in the population depending on the regarded city, the non-epidemic indicence baseline, the contact structure considered and the used calibration method. The experiments demonstrated the importance of distinguishing characteristics of different age groups, such as contact intensities and background immunity levels. It was also found that the current model calibration process, which relies mostly on ARI incidence, demonstrates notable variation of output parameter values. Employing additional sources of data, such as strain-specific influenza incidence and external assessments on underreporting levels in different age groups, might enhance the plausibility of parameter values obtained by model calibration, along with reducing the assessment variation.

Genome-wide characterization of the seasonal H3N2 virus in Shanghai reveals natural temperature-sensitive strains conferred by the I668V mutation in the PA subunit

Seasonal H3N2 influenza viruses are recognized as major epidemic viruses, exhibiting complex seasonal patterns in regions with temperate climates. To investigate the influence of viral evolution and mutations on the seasonality of influenza, we performed a genome-wide analysis of samples collected from 62 influenza A/H3N2-infected patients in Shanghai during 2016-2017. Phylogenetic analysis of all eight segments of the influenza A virus revealed that there were two epidemic influenza virus strains circulating in the 2016-2017 winter season (2016-2017win) and 2017 summer season (2017sum). Replication of the two epidemic viral strains at different temperatures (33, 35, 37, and 39 °C) was measured, and the correlation of the mutations in the two epidemic viral strains with temperature sensitivity and viral replication was analyzed. Analysis of the replication kinetics showed that replication of the 2016-2017win strains was significantly restricted at 39 °C compared with that of the 2017sum strains. A polymerase activity assay and mutational analysis demonstrated that the PA I668V mutation of the 2016-2017win viruses suppressed polymerase activity in vitro at high temperatures. Taken together, these data suggest that the I668V mutation in the PA subunit of the 2016-2017win strains may confer temperature sensitivity and attenuate viral replication and polymerase activity; meanwhile, the 2017sum strains maintained virulence at high temperatures. These findings highlight the importance of certain mutations in viral adaptation and persistence in subsequent seasons.

Accounting for non-stationarity in epidemiology by embedding time-varying parameters in stochastic models

The spread of disease through human populations is complex. The characteristics of disease propagation evolve with time, as a result of a multitude of environmental and anthropic factors, this non-stationarity is a key factor in this huge complexity. In the absence of appropriate external data sources, to correctly describe the disease propagation, we explore a flexible approach, based on stochastic models for the disease dynamics, and on diffusion processes for the parameter dynamics. Using such a diffusion process has the advantage of not requiring a specific mathematical function for the parameter dynamics. Coupled with particle MCMC, this approach allows us to reconstruct the time evolution of some key parameters (average transmission rate for instance). Thus, by capturing the time-varying nature of the different mechanisms involved in disease propagation, the epidemic can be described. Firstly we demonstrate the efficiency of this methodology on a toy model, where the parameters and the observation process are known. Applied then to real datasets, our methodology is able, based solely on simple stochastic models, to reconstruct complex epidemics, such as flu or dengue, over long time periods. Hence we demonstrate that time-varying parameters can improve the accuracy of model performances, and we suggest that our methodology can be used as a first step towards a better understanding of a complex epidemic, in situation where data is limited and/or uncertain.

Evolution-informed forecasting of seasonal influenza A (H3N2)

Interpandemic or seasonal influenza A, currently subtypes H3N2 and H1N1, exacts an enormous annual burden both in terms of human health and economic impact. Incidence prediction ahead of season remains a challenge largely because of the virus' antigenic evolution. We propose a forecasting approach that incorporates evolutionary change into a mechanistic epidemiological model. The proposed models are simple enough that their parameters can be estimated from retrospective surveillance data. These models link amino acid sequences of hemagglutinin epitopes with a transmission model for seasonal H3N2 influenza, also informed by H1N1 levels. With a monthly time series of H3N2 incidence in the United States for more than 10 years, we demonstrate the feasibility of skillful prediction for total cases ahead of season, with a tendency to underpredict monthly peak epidemic size, and an accurate real-time forecast for the 2016/2017 influenza season.

A switching model for the impact of toxins on the spread of infectious diseases

To study the effects of an environmental toxin, such as fine particles in hazy weather, on the spread of infectious diseases, we derive a toxin-dependent dynamic model that incorporates the birth rate with the toxin-dependent switching mode, the mortality rate, and infection rate with the toxin-dependent saturation effect. We analyze the model by showing the positive invariance, existence and stability of equilibria, and bifurcations. Numerical simulation is adopted to verify the mathematical results and exhibit transcritical and Hopf bifurcations. Our theoretical results show that there exists a threshold value of the environmental toxin: if the environmental toxin concentration is lower than the threshold, the system has a disease-free equilibrium and an interior equilibrium; if the environmental toxin concentration is higher than the threshold, the system has the extinction equilibrium. For the case where the disease-induced death is ignored, we show the global stability results. Numerical simulations clearly show that the environmental toxin facilitates the spread of infectious diseases. This study provides a theoretical basis for uncovering the impact of toxins on the spread of infectious diseases and for guiding the decision making by disease control agencies and governments.

Prediction of influenza peaks in Russian cities: Comparing the accuracy of two SEIR models

This paper is dedicated to the application of two types of SEIR models to the influenza outbreak peak prediction in Russian cities. The first one is a continuous SEIR model described by a system of ordinary differential equations. The second one is a discrete model formulated as a set of difference equations, which was used in the Baroyan-Rvachev modeling framework for the influenza outbreak prediction in the Soviet Union. The outbreak peak day and height predictions were performed by calibrating both models to varied-size samples of long-term data on ARI incidence in Moscow, Saint Petersburg, and Novosibirsk. The accuracy of the modeling predictions on incomplete data was compared with a number of other peak forecasting methods tested on the same dataset. The drawbacks of the described prediction approach and possible ways to overcome them are discussed.

A two-stage approach for estimating the parameters of an age-group epidemic model from incidence data

Age-dependent dynamics is an important characteristic of many infectious diseases. Age-group epidemic models describe the infection dynamics in different age-groups by allowing to set distinct parameter values for each. However, such models are highly nonlinear and may have a large number of unknown parameters. Thus, parameter estimation of age-group models, while becoming a fundamental issue for both the scientific study and policy making in infectious diseases, is not a trivial task in practice. In this paper, we examine the estimation of the so-called next-generation matrix using incidence data of a single entire outbreak, and extend the approach to deal with recurring outbreaks. Unlike previous studies, we do not assume any constraints regarding the structure of the matrix. A novel two-stage approach is developed, which allows for efficient parameter estimation from both statistical and computational perspectives. Simulation studies corroborate the ability to estimate accurately the parameters of the model for several realistic scenarios. The model and estimation method are applied to real data of influenza-like-illness in Israel. The parameter estimates of the key relevant epidemiological parameters and the recovered structure of the estimated next-generation matrix are in line with results obtained in previous studies.

The use of ambient humidity conditions to improve influenza forecast

Laboratory and epidemiological evidence indicate that ambient humidity modulates the survival and transmission of influenza. Here we explore whether the inclusion of humidity forcing in mathematical models describing influenza transmission improves the accuracy of forecasts generated with those models. We generate retrospective forecasts for 95 cities over 10 seasons in the United States and assess both forecast accuracy and error. Overall, we find that humidity forcing improves forecast performance (at 1-4 lead weeks, 3.8% more peak week and 4.4% more peak intensity forecasts are accurate than with no forcing) and that forecasts generated using daily climatological humidity forcing generally outperform forecasts that utilize daily observed humidity forcing (4.4% and 2.6% respectively). These findings hold for predictions of outbreak peak intensity, peak timing, and incidence over 2- and 4-week horizons. The results indicate that use of climatological humidity forcing is warranted for current operational influenza forecast.

Modelling and parameter inference of predator-prey dynamics in heterogeneous environments using the direct integral approach

Most bacterial habitats are topographically complex in the micro scale. Important examples include the gastrointestinal and tracheal tracts, and the soil. Although there are myriad theoretical studies that explore the role of spatial structures on antagonistic interactions (predation, competition) among animals, there are many fewer experimental studies that have explored, validated and quantified their predictions. In this study, we experimentally monitored the temporal dynamic of the predatory bacterium Bdellovibrio bacteriovorus, and its prey, the bacterium Burkholderia stabilis in a structured habitat consisting of sand under various regimes of wetness. We constructed a dynamic model, and estimated its parameters by further developing the direct integral method, a novel estimation procedure that exploits the separability of the states and parameters in the model. We also verified that one of our parameter estimates was consistent with its known, directly measured value from the literature. The ability of the model to fit the data combined with realistic parameter estimates indicate that bacterial predation in the sand can be described by a relatively simple model, and stress the importance of prey refuge on predation dynamics in heterogeneous environments.

Influenza Season Hospitalization Trends in Israel: A Multi-Year Comparative Analysis 2005/2006 Through 2012/2013

BACKGROUND

Influenza-related morbidity impacts healthcare systems, including hospitals.

OBJECTIVE

To obtain a quantitative assessment of hospitalization burden in pediatric and internal medicine departments during influenza seasons compared with the summer months in Israel.

METHODS

Data on pediatric and internal medicine hospitalized patients in general hospitals in Israel during the influenza seasons between 2005 and 2013 were analyzed for rate of hospitalizations, rate of hospitalization days, hospital length of stay (LOS), and bed occupancy and compared with the summer months. Data were analyzed for hospitalizations for all diagnoses, diagnoses of respiratory or cardiovascular disease (ICD9 390-519), and influenza or pneumonia (ICD9480-487), with data stratified by age. The 2009-2010 pandemic influenza season was excluded.

RESULTS

Rates of monthly hospitalizations and hospitalization days for all diagnoses were 4.8% and 8% higher, respectively, during influenza seasons as compared with the summers. The mean LOS per hospitalization for all diagnoses demonstrated a small increase during influenza seasons as compared with summer seasons. The excess hospitalizations and hospitalization days were especially noticed for the age groups under 1 year, 1-4 years, and 85 years and older. The differences were severalfold higher for patients with a diagnosis of respiratory or cardiovascular disease and influenza or pneumonia. Bed occupancy was higher during influenza seasons compared with the summer, particularly in pediatric departments.

CONCLUSIONS

Hospital burden in pediatric and internal medicine departments during influenza seasons in Israel was associated with age and diagnosis. These results are important for optimal preparedness for influenza seasons.

Indoor temperature and humidity in New York City apartments during winter

BACKGROUND

Concerns about indoor residential humidity have largely centered on dampness prevention. Overly dry air, however, may favor the survival of some viruses and hence respiratory infections. Many residents employ portable humidifiers to humidify their home environment, yet the effect of these humidifiers on indoor humidity is not known.

METHODS

We monitored indoor temperature and humidity in 34 apartments in New York City during winter 2014-2015. We combined information from the monitors with surveyed information on building, household, and apartment-level factors and with information on household humidifier use. Using multilevel regression models, we investigated the role of these factors on indoor absolute humidity levels during the winter.

RESULTS

Mean indoor vapor pressure (a measure of absolute humidity) was 6.7mb in the surveyed homes during the winter season. Ownership of a humidifier was not associated with higher indoor humidity levels; however, larger building size (above 100units) was significantly associated with lower humidity. The presence of a radiator heating system was non-significantly associated with higher humidity.

CONCLUSIONS

The wintertime indoor environment in this sample of New York City apartments is dry. Future research is needed to evaluate the effectiveness of portable humidifiers in the home and to clarify the relationship between dry indoor air and the transmission of viral infections.

Model selection for seasonal influenza forecasting

Epidemics of seasonal influenza inflict a huge burden in temperate climes such as Melbourne (Australia) where there is also significant variability in their timing and magnitude. Particle filters combined with mechanistic transmission models for the spread of influenza have emerged as a popular method for forecasting the progression of these epidemics. Despite extensive research it is still unclear what the optimal models are for forecasting influenza, and how one even measures forecast performance.

In this paper, we present a likelihood-based method, akin to Bayes factors, for model selection when the aim is to select for predictive skill. Here, “predictive skill” is measured by the probability of the data after the forecasting date, conditional on the data from before the forecasting date. Using this method we choose an optimal model of influenza transmission to forecast the number of laboratory-confirmed cases of influenza in Melbourne in each of the 2010–15 epidemics. The basic transmission model considered has the susceptible-exposed-infectious-recovered structure with extensions allowing for the effects of absolute humidity and inhomogeneous mixing in the population. While neither of the extensions provides a significant improvement in fit to the data they do differ in terms of their predictive skill. Both measurements of absolute humidity and a sinusoidal approximation of those measurements are observed to increase the predictive skill of the forecasts, while allowing for inhomogeneous mixing reduces the skill.

We discuss how our work could be integrated into a forecasting system and how the model selection method could be used to evaluate forecasts when comparing to multiple surveillance systems providing disparate views of influenza activity.

Model-based reconstruction of an epidemic using multiple datasets: understanding influenza A/H1N1 pandemic dynamics in Israel

Intensified surveillance during the 2009 A/H1N1 influenza pandemic in Israel resulted in large virological and serological datasets, presenting a unique opportunity for investigating the pandemic dynamics. We employ a conditional likelihood approach for fitting a disease transmission model to virological and serological data, conditional on clinical data. The model is used to reconstruct the temporal pattern of the pandemic in Israel in five age-groups and evaluate the factors that shaped it. We estimate the reproductive number at the beginning of the pandemic to beR= 1.4. We find that the combined effect of varying absolute humidity conditions and school vacations (SVs) is responsible for the infection pattern, characterized by three epidemic waves. Overall attack rate is estimated at 32% (28-35%) with a large variation among the age-groups: the highest attack rates within school children and the lowest within the elderly. This pattern of infection is explained by a combination of the age-group contact structure and increasing immunity with age. We assess that SVs increased the overall attack rates by prolonging the pandemic into the winter. Vaccinating school children would have been the optimal strategy for minimizing infection rates in all age-groups.

Predicting Fluctuating Rates of Hospitalizations in Relation to Influenza Epidemics and Meteorological Factors

Introduction

In France, rates of hospital admissions increase at the peaks of influenza epidemics. Predicting influenza-associated hospitalizations could help to anticipate increased hospital activity. The purpose of this study is to identify predictors of influenza epidemics through the analysis of meteorological data, and medical data provided by general practitioners.

Methods

Historical data were collected from Meteo France, the Sentinelles network and hospitals’ information systems for a period of 8 years (2007–2015). First, connections between meteorological and medical data were estimated with the Pearson correlation coefficient, Principal component analysis and classification methods (Ward and k-means). Epidemic states of tested weeks were then predicted for each week during a one-year period using linear discriminant analysis. Finally, transition probabilities between epidemic states were calculated with the Markov Chain method.

Results

High correlations were found between influenza-associated hospitalizations and the variables: Sentinelles and emergency department admissions, and anti-correlations were found between hospitalizations and each of meteorological factors applying a time lag of: -13, -12 and -32 days respectively for temperature, absolute humidity and solar radiation. Epidemic weeks were predicted accurately with the linear discriminant analysis method; however there were many misclassifications about intermediate and non-epidemic weeks. Transition probability to an epidemic state was 100% when meteorological variables were below: 2°C, 4 g/m³ and 32 W/m², respectively for temperature, absolute humidity and solar radiation. This probability was 0% when meteorological variables were above: 6°C, 5.8g/m³ and 74W/m².

Conclusion

These results confirm a good correlation between influenza-associated hospitalizations, meteorological factors and general practitioner’s activity, the latter being the strongest predictor of hospital activity.

Quantifying the role of weather on seasonal influenza

BACKGROUND

Improving knowledge about influenza transmission is crucial to upgrade surveillance network and to develop accurate predicting models to enhance public health intervention strategies. Epidemics usually occur in winter in temperate countries and during the rainy season for tropical countries, suggesting a climate impact on influenza spread. Despite a lot of studies, the role of weather on influenza spread is not yet fully understood. In the present study, we investigated this issue at two different levels.

METHODS

First, we evaluated how weekly (intra-annual) incidence variations of clinical diseases could be linked to those of climatic factors. We considered that only a fraction of the human population is susceptible at the beginning of a year due to immunity acquired from previous years. Second, we focused on epidemic sizes (cumulated number of clinical reported cases) and looked at how their inter-annual and regional variations could be related to differences in the winter climatic conditions of the epidemic years over the regions. We quantified the impact of fifteen climatic variables in France using the Réseau des GROG surveillance network incidence data over eleven regions and nine years.

RESULTS

At the epidemic scale, no impact of climatic factors was highlighted. At the intra-annual scale, six climatic variables had a significant impact: average temperature (5.54 ± 1.09 %), absolute humidity (5.94 ± 1.08 %), daily variation of absolute humidity (3.02 ± 1.17 %), sunshine duration (3.46 ± 1.06 %), relative humidity (4.92 ± 1.20 %) and daily variation of relative humidity (4.46 ± 1.24 %). Since in practice the impact of two highly correlated variables is very hard to disentangle, we performed a principal component analysis that revealed two groups of three highly correlated climatic variables: one including the first three highlighted climatic variables on the one hand, the other including the last three ones on the other hand.

CONCLUSIONS

These results suggest that, among the six factors that appeared to be significant, only two (one per group) could in fact have a real effect on influenza spread, although it is not possible to determine which one based on a purely statistical argument. Our results support the idea of an important role of climate on the spread of influenza.

Humidity: A review and primer on atmospheric moisture and human health

Research examining associations between weather and human health frequently includes the effects of atmospheric humidity. A large number of humidity variables have been developed for numerous purposes, but little guidance is available to health researchers regarding appropriate variable selection. We examine a suite of commonly used humidity variables and summarize both the medical and biometeorological literature on associations between humidity and human health. As an example of the importance of humidity variable selection, we correlate numerous hourly humidity variables to daily respiratory syncytial virus isolates in Singapore from 1992 to 1994. Most water-vapor mass based variables (specific humidity, absolute humidity, mixing ratio, dewpoint temperature, vapor pressure) exhibit comparable correlations. Variables that include a thermal component (relative humidity, dewpoint depression, saturation vapor pressure) exhibit strong diurnality and seasonality. Humidity variable selection must be dictated by the underlying research question. Despite being the most commonly used humidity variable, relative humidity should be used sparingly and avoided in cases when the proximity to saturation is not medically relevant. Care must be taken in averaging certain humidity variables daily or seasonally to avoid statistical biasing associated with variables that are inherently diurnal through their relationship to temperature.

Associations between Meteorological Parameters and Influenza Activity in Berlin (Germany), Ljubljana (Slovenia), Castile and León (Spain) and Israeli Districts

BACKGROUND

Studies in the literature have indicated that the timing of seasonal influenza epidemic varies across latitude, suggesting the involvement of meteorological and environmental conditions in the transmission of influenza. In this study, we investigated the link between meteorological parameters and influenza activity in 9 sub-national areas with temperate and subtropical climates: Berlin (Germany), Ljubljana (Slovenia), Castile and León (Spain) and all 6 districts in Israel.

METHODS

We estimated weekly influenza-associated influenza-like-illness (ILI) or Acute Respiratory Infection (ARI) incidence to represent influenza activity using data from each country's sentinel surveillance during 2000-2011 (Spain) and 2006-2011 (all others). Meteorological data was obtained from ground stations, satellite and assimilated data. Two generalized additive models (GAM) were developed, with one using specific humidity as a covariate and another using minimum temperature. Precipitation and solar radiation were included as additional covariates in both models. The models were adjusted for previous weeks' influenza activity, and were trained separately for each study location.

RESULTS

Influenza activity was inversely associated (p<0.05) with specific humidity in all locations. Minimum temperature was inversely associated with influenza in all 3 temperate locations, but not in all subtropical locations. Inverse associations between influenza and solar radiation were found in most locations. Associations with precipitation were location-dependent and inconclusive. We used the models to estimate influenza activity a week ahead for the 2010/2011 period which was not used in training the models. With exception of Ljubljana and Israel's Haifa District, the models could closely follow the observed data especially during the start and the end of epidemic period. In these locations, correlation coefficients between the observed and estimated ranged between 0.55 to 0.91and the model-estimated influenza peaks were within 3 weeks from the observations.

CONCLUSION

Our study demonstrated the significant link between specific humidity and influenza activity across temperate and subtropical climates, and that inclusion of meteorological parameters in the surveillance system may further our understanding of influenza transmission patterns.

Global Spatio-temporal Patterns of Influenza in the Post-pandemic Era

We study the global spatio-temporal patterns of influenza dynamics. This is achieved by analysing and modelling weekly laboratory confirmed cases of influenza A and B from 138 countries between January 2006 and January 2015. The data were obtained from FluNet, the surveillance network compiled by the the World Health Organization. We report a pattern of skip-and-resurgence behavior between the years 2011 and 2013 for influenza H1N1pdm, the strain responsible for the 2009 pandemic, in Europe and Eastern Asia. In particular, the expected H1N1pdm epidemic outbreak in 2011/12 failed to occur (or "skipped") in many countries across the globe, although an outbreak occurred in the following year. We also report a pattern of well-synchronized wave of H1N1pdm in early 2011 in the Northern Hemisphere countries, and a pattern of replacement of strain H1N1pre by H1N1pdm between the 2009 and 2012 influenza seasons. Using both a statistical and a mechanistic mathematical model, and through fitting the data of 108 countries, we discuss the mechanisms that are likely to generate these events taking into account the role of multi-strain dynamics. A basic understanding of these patterns has important public health implications and scientific significance.

Multiannual forecasting of seasonal influenza dynamics reveals climatic and evolutionary drivers

Human influenza occurs annually in most temperate climatic zones of the world, with epidemics peaking in the cold winter months. Considerable debate surrounds the relative role of epidemic dynamics, viral evolution, and climatic drivers in driving year-to-year variability of outbreaks. The ultimate test of understanding is prediction; however, existing influenza models rarely forecast beyond a single year at best. Here, we use a simple epidemiological model to reveal multiannual predictability based on high-quality influenza surveillance data for Israel; the model fit is corroborated by simple metapopulation comparisons within Israel. Successful forecasts are driven by temperature, humidity, antigenic drift, and immunity loss. Essentially, influenza dynamics are a balance between large perturbations following significant antigenic jumps, interspersed with nonlinear epidemic dynamics tuned by climatic forcing.

Mathematical modelling and prediction in infectious disease epidemiology

We discuss to what extent disease transmission models provide reliable predictions. The concept of prediction is delineated as it is understood by modellers, and illustrated by some classic and recent examples. A precondition for a model to provide valid predictions is that the assumptions underlying it correspond to the reality, but such correspondence is always limited—all models are simplifications of reality. A central tenet of the modelling enterprise is what we may call the ‘robustness thesis’: a model whose assumptions approximately correspond to reality will make predictions that are approximately valid. To examine which of the predictions made by a model are trustworthy, it is essential to examine the outcomes of different models. Thus, if a highly simplified model makes a prediction, and if the same or a very similar prediction is made by a more elaborate model that includes some mechanisms or details that the first model did not, then we gain some confidence that the prediction is robust. An important benefit derived from mathematical modelling activity is that it demands transparency and accuracy regarding our assumptions, thus enabling us to test our understanding of the disease epidemiology by comparing model results and observed patterns. Models can also assist in decision-making by making projections regarding important issues such as intervention-induced changes in the spread of disease.

Spatio-temporal synchrony of influenza in cities across Israel: the "Israel is one city" hypothesis

We analysed an 11-year dataset (1998-2009) of Influenza-Like Illness (ILI) that was based on surveillance of ∽23% of Israel's population. We examined whether the level of synchrony of ILI epidemics in Israel's 12 largest cities is high enough to view Israel as a single epidemiological unit. Two methods were developed to assess the synchrony: (1) City-specific attack rates were fitted to a simple model in order to estimate the temporal differences in attack rates and spatial differences in reporting rates of ILI. The model showed good fit to the data (R2  =  0.76) and revealed considerable differences in reporting rates of ILI in different cities (up to a factor of 2.2). (2) A statistical test was developed to examine the null hypothesis (H0) that ILI incidence curves in two cities are essentially identical, and was tested using ILI data. Upon examining all possible pairs of incidence curves, 77.4% of pairs were found not to be different (H0 was not rejected). It was concluded that all cities generally have the same attack rate and follow the same epidemic curve each season, although the attack rate changes from season to season, providing strong support for the "Israel is one city" hypothesis. The cities which were the most out of synchronization were Bnei Brak, Beersheba and Haifa, the latter two being geographically remote from all other cities in the dataset and the former geographically very close to several other cities but socially separate due to being populated almost exclusively by ultra-orthodox Jews. Further evidence of assortative mixing of the ultra-orthodox population can be found in the 2001-2002 season, when ultra-orthodox cities and neighborhoods showed distinctly different incidence curves compared to the general population.