Search

The pathogenic bacteria Neisseria meningitidis, which causes invasive meningococcal disease (IMD), predominantly colonizes humans asymptomatically; however, invasive disease occurs in a small proportion of the population. Here, we explore the seasonality of IMD and develop and validate a suite of models for simulating and forecasting disease outcomes in the United States. We combine the models into multi-model ensembles (MME) based on the past performance of the individual models, as well as a naive equally weighted aggregation, and compare the retrospective forecast performance over a six-month forecast horizon. Deployment of the complete vaccination regimen, introduced in 2011, coincided with a change in the periodicity of IMD, suggesting altered transmission dynamics. We found that a model forced with the period obtained by local power wavelet decomposition best fit and forecast observations. In addition, the MME performed the best across the entire study period. Finally, our study included US-level data until 2022, allowing study of a possible IMD rebound after relaxation of non-pharmaceutical interventions imposed in response to the COVID-19 pandemic; however, no evidence of a rebound was found. Our findings demonstrate the ability of process-based models to retrospectively forecast IMD and provide a first analysis of the seasonality of IMD before and after the complete vaccination regimen. Author summary: This paper presents a time series analysis of Invasive Meningococcal Disease caused by the pathogenic bacteria Neisseria meningitidis as well as the development and validation of a suite of models for simulating and forecasting disease outcomes in the United States. We present the models and their performance and construct a multi-model ensemble (MME) forecast that combines individual models based on past forecast performance, as well as an equally weighted aggregation. Deployment of a complete vaccination regimen, introduced for adolescents in 2011, coincided with a change in the periodicity of IMD, suggesting altered transmission dynamics. We found that a model forced with the period obtained by local power wavelet decomposition best fits and forecasts observations. In addition, the MME forecasts performed the best across the entire study period. Finally, our study included US-level data until 2022, allowing study of a possible IMD rebound after the relaxation of non-pharmaceutical interventions imposed in response to the COVID-19 pandemic; however, no evidence of a rebound was found. Our findings demonstrate the ability of process-based models to retrospectively forecast IMD and provide a first analysis of the seasonality of IMD before and after the complete vaccination regimen. [ABSTRACT FROM AUTHOR]

Influenza affects millions of people every year. It causes a considerable amount of medical visits and hospitalisations as well as hundreds of thousands of deaths. Forecasting influenza prevalence with good accuracy can significantly help public health agencies to timely react to seasonal or novel strain epidemics. Although significant progress has been made, influenza forecasting remains a challenging modelling task. In this paper, we propose a methodological framework that improves over the state-of-the-art forecasting accuracy of influenza-like illness (ILI) rates in the United States. We achieve this by using Web search activity time series in conjunction with historical ILI rates as observations for training neural network (NN) architectures. The proposed models incorporate Bayesian layers to produce associated uncertainty intervals to their forecast estimates, positioning themselves as legitimate complementary solutions to more conventional approaches. The best performing NN, referred to as the iterative recurrent neural network (IRNN) architecture, reduces mean absolute error by 10.3% and improves skill by 17.1% on average in nowcasting and forecasting tasks across 4 consecutive flu seasons. Author summary: Modelling the prevalence of an infectious disease enables public health organisations to prepare for and minimise its impact. However, traditional disease indicators are often quite restrictive as they provide information with a significant delay. Recent research efforts have provided evidence of the value of alternative information sources such as Web search activity trends. Our work incorporates this information into influenza forecasting models to achieve state-of-the-art accuracy. In addition, the proposed Bayesian neural network architectures also provide associated uncertainty estimates for the forecasts, positioning our methodology as a practical complementary tool for disease surveillance and policy making. [ABSTRACT FROM AUTHOR]

Measles is one the best-documented and most-mechanistically-studied non-linear infectious disease dynamical systems. However, systematic investigation into the comparative performance of traditional mechanistic models and machine learning approaches in forecasting the transmission dynamics of this pathogen are still rare. Here, we compare one of the most widely used semi-mechanistic models for measles (TSIR) with a commonly used machine learning approach (LASSO), comparing performance and limits in predicting short to long term outbreak trajectories and seasonality for both regular and less regular measles outbreaks in England and Wales (E&W) and the United States. First, our results indicate that the proposed LASSO model can efficiently use data from multiple major cities and achieve similar short-to-medium term forecasting performance to semi-mechanistic models for E&W epidemics. Second, interestingly, the LASSO model also captures annual to biennial bifurcation of measles epidemics in E&W caused by susceptible response to the late 1940s baby boom. LASSO may also outperform TSIR for predicting less-regular dynamics such as those observed in major cities in US between 1932–45. Although both approaches capture short-term forecasts, accuracy suffers for both methods as we attempt longer-term predictions in highly irregular, post-vaccination outbreaks in E&W. Finally, we illustrate that the LASSO model can both qualitatively and quantitatively reconstruct mechanistic assumptions, notably susceptible dynamics, in the TSIR model. Our results characterize the limits of predictability of infectious disease dynamics for strongly immunizing pathogens with both mechanistic and machine learning models, and identify connections between these two approaches. Author summary: Machine learning techniques in infectious disease modeling have grown in popularity in recent years. However, systematic investigation into the comparative performance of these approaches with traditional mechanistic models are still rare. In this paper, we compare one of the most widely used semi-mechanistic models for measles (TSIR) with a commonly used machine learning approach (LASSO), comparing performance and limits in predicting short to long term outbreaks of measles, one of the best-documented and most-mechanistically-studied non-linear infectious disease dynamical systems. Our results show that in general the LASSO outperform TSIR for predicting less-regular dynamics, and it can achieve similar performance in other scenarios when compared to the TSIR. The LASSO also has the advantages of not requiring explicit demographic data in model training. Finally, we identify connections between these two approaches and show that the LASSO model can both qualitatively and quantitatively reconstruct mechanistic assumptions, notably susceptible dynamics, in the TSIR model. [ABSTRACT FROM AUTHOR]

Reported COVID-19 cases and deaths provide a delayed and incomplete picture of SARS-CoV-2 infections in the United States (US). Accurate estimates of both the timing and magnitude of infections are needed to characterize viral transmission dynamics and better understand COVID-19 disease burden. We estimated time trends in SARS-CoV-2 transmission and other COVID-19 outcomes for every county in the US, from the first reported COVID-19 case in January 13, 2020 through January 1, 2021. To do so we employed a Bayesian modeling approach that explicitly accounts for reporting delays and variation in case ascertainment, and generates daily estimates of incident SARS-CoV-2 infections on the basis of reported COVID-19 cases and deaths. The model is freely available as the covidestim R package. Nationally, we estimated there had been 49 million symptomatic COVID-19 cases and 404,214 COVID-19 deaths by the end of 2020, and that 28% of the US population had been infected. There was county-level variability in the timing and magnitude of incidence, with local epidemiological trends differing substantially from state or regional averages, leading to large differences in the estimated proportion of the population infected by the end of 2020. Our estimates of true COVID-19 related deaths are consistent with independent estimates of excess mortality, and our estimated trends in cumulative incidence of SARS-CoV-2 infection are consistent with trends in seroprevalence estimates from available antibody testing studies. Reconstructing the underlying incidence of SARS-CoV-2 infections across US counties allows for a more granular understanding of disease trends and the potential impact of epidemiological drivers. Author summary: Because many COVID-19 infections go undetected, reported numbers of COVID-19 cases and deaths underestimate the true size of the epidemic. To address this problem, we built a model to estimate the number of new SARS-CoV-2 infections over time in each U.S. state and county. In this paper, we present time trends of infections and other disease outcomes from the first reported case in the U.S. until January 1, 2021, for each state and county. The time series of infection estimates suggest that the US epidemic is best described a series of related epidemics, varying in their timing and intensity. We estimate that over a quarter of the US population was infected with SARS-CoV-2 in 2020 and by the end of 2020 0.12% of the population had died from COVID-19. State-level results were consistent with external measures of disease burden, including estimates of SARS-CoV-2 seroprevalence and excess mortality. Our findings help better understand the epidemic in the pre-vaccine era and demonstrate the feasibility of estimating SARS-CoV-2 infections at local levels using routinely reported case and death data. [ABSTRACT FROM AUTHOR]

Prophylactic interventions such as vaccine allocation are some of the most effective public health policy planning tools. The supply of vaccines, however, is limited and an important challenge is to optimally allocate the vaccines to minimize epidemic impact. This resource allocation question (which we refer to as VID) has multiple dimensions: when, where, to whom, etc. Most of the existing literature in this topic deals with the latter (to whom), proposing policies that prioritize individuals by age and disease risk. However, since seasonal influenza spread has a typical spatial trend, and due to the temporal constraints enforced by the availability schedule, the when and where problems become equally, if not more, relevant. In this paper, we study the VID problem in the context of seasonal influenza spread in the United States. We develop a national scale metapopulation model for influenza that integrates both short and long distance human mobility, along with realistic data on vaccine uptake. We also design GA, a greedy algorithm for allocating the vaccine supply at the state level under temporal constraints and show that such a strategy improves over the current baseline of pro-rata allocation, and the improvement is more pronounced for higher vaccine efficacy and moderate flu season intensity. Further, the resulting strategy resembles a ring vaccination applied spatiallyacross the US. [ABSTRACT FROM AUTHOR]

The ability to produce timely and accurate flu forecasts in the United States can significantly impact public health. Augmenting forecasts with internet data has shown promise for improving forecast accuracy and timeliness in controlled settings, but results in practice are less convincing, as models augmented with internet data have not consistently outperformed models without internet data. In this paper, we perform a controlled experiment, taking into account data backfill, to improve clarity on the benefits and limitations of augmenting an already good flu forecasting model with internet-based nowcasts. Our results show that a good flu forecasting model can benefit from the augmentation of internet-based nowcasts in practice for all considered public health-relevant forecasting targets. The degree of forecast improvement due to nowcasting, however, is uneven across forecasting targets, with short-term forecasting targets seeing the largest improvements and seasonal targets such as the peak timing and intensity seeing relatively marginal improvements. The uneven forecasting improvements across targets hold even when “perfect” nowcasts are used. These findings suggest that further improvements to flu forecasting, particularly seasonal targets, will need to derive from other, non-nowcasting approaches. [ABSTRACT FROM AUTHOR]

Ancestral state reconstruction in Bayesian phylogeography of virus pandemics have been improved by utilizing a Bayesian stochastic search variable selection (BSSVS) framework. Recently, this framework has been extended to model the transition rate matrix between discrete states as a generalized linear model (GLM) of genetic, geographic, demographic, and environmental predictors of interest to the virus and incorporating BSSVS to estimate the posterior inclusion probabilities of each predictor. Although the latter appears to enhance the biological validity of ancestral state reconstruction, there has yet to be a comparison of phylogenies created by the two methods. In this paper, we compare these two methods, while also using a primitive method without BSSVS, and highlight the differences in phylogenies created by each. We test six coalescent priors and six random sequence samples of H3N2 influenza during the 2014–15 flu season in the U.S. We show that the GLMs yield significantly greater root state posterior probabilities than the two alternative methods under five of the six priors, and significantly greater Kullback-Leibler divergence values than the two alternative methods under all priors. Furthermore, the GLMs strongly implicate temperature and precipitation as driving forces of this flu season and nearly unanimously identified a single root state, which exhibits the most tropical climate during a typical flu season in the U.S. The GLM, however, appears to be highly susceptible to sampling bias compared with the other methods, which casts doubt on whether its reconstructions should be favored over those created by alternate methods. We report that a BSSVS approach with a Poisson prior demonstrates less bias toward sample size under certain conditions than the GLMs or primitive models, and believe that the connection between reconstruction method and sampling bias warrants further investigation. [ABSTRACT FROM AUTHOR]