RESUMEN
Mathematical models in ecology and epidemiology must be consistent with observed data in order to generate reliable knowledge and evidence-based policy. Metapopulation systems, which consist of a network of connected sub-populations, pose technical challenges in statistical inference owing to nonlinear, stochastic interactions. Numerical difficulties encountered in conducting inference can obstruct the core scientific questions concerning the link between the mathematical models and the data. Recently, an algorithm has been proposed that enables computationally tractable likelihood-based inference for high-dimensional partially observed stochastic dynamic models of metapopulation systems. We use this algorithm to build a statistically principled data analysis workflow for metapopulation systems. Via a case study of COVID-19, we show how this workflow addresses the limitations of previous approaches. The COVID-19 pandemic provides a situation where mathematical models and their policy implications are widely visible, and we revisit an influential metapopulation model used to inform basic epidemiological understanding early in the pandemic. Our methods support self-critical data analysis, enabling us to identify and address model weaknesses, leading to a new model with substantially improved statistical fit and parameter identifiability. Our results suggest that the lockdown initiated on 23 January 2020 in China was more effective than previously thought.
Asunto(s)
COVID-19 , SARS-CoV-2 , COVID-19/epidemiología , Humanos , Algoritmos , Modelos Biológicos , Dinámica Poblacional , PandemiasRESUMEN
We consider genealogies arising from a Markov population process in which individuals are categorized into a discrete collection of compartments, with the requirement that individuals within the same compartment are statistically exchangeable. When equipped with a sampling process, each such population process induces a time-evolving tree-valued process defined as the genealogy of all sampled individuals. We provide a construction of this genealogy process and derive exact expressions for the likelihood of an observed genealogy in terms of filter equations. These filter equations can be numerically solved using standard Monte Carlo integration methods. Thus, we obtain statistically efficient likelihood-based inference for essentially arbitrary compartment models based on an observed genealogy of individuals sampled from the population.
RESUMEN
Bacterial pathogens that are successful in hospital environments must survive times of intense antibiotic exposure and times of no antibiotic exposure. When these organisms are closely associated with human hosts, they must also transmit from one patient to another for the resistance to spread. The resulting evolutionary dynamics have, in some settings, led to rising levels of resistance in hospitals. Here, we focus on an important but understudied aspect of this dynamic: the loss of resistance when the resistant organisms evolve in environments where the antibiotic pressure is removed. Based on prior data, we hypothesize that resistance arising in the context of strong selection may carry a high cost and revert to sensitivity quickly once the selective pressure is removed. Conversely, resistant isolates that persist through times of no antibiotic pressure should carry a lower cost and revert less quickly. To test this hypothesis, we utilize a genetically diverse set of patient-derived, daptomycin-resistant Enterococcus faecium isolates that include cases of both de novo emergence of resistance within patients and putatively transmitted resistance. Both of these sets of strains have survived periods of antibiotic exposure, but only putatively transmitted resistant strains have survived extended periods without antibiotic exposure. These strains were then allowed to evolve in antibiotic free laboratory conditions. We find that putatively transmitted resistant strains tended to have lower level resistance but that evolution in antibiotic-free conditions resulted in minimal loss of resistance. In contrast, resistance that arose de novo within patients was higher level but exhibited greater declines in resistance in vitro. Sequencing of the experimentally evolved isolates revealed that reversal of high level resistance resulted from evolutionary pathways that were frequently genetically associated with the unique resistance mutations of that strain. Thus, the rapid reversal of high-level resistance was associated with accessible evolutionary pathways where an increase in fitness is associated with decreased resistance. We describe how this rapid loss of resistance may limit the spread of resistance within the hospital and shape the diversity of resistance phenotypes across patients.
RESUMEN
Reassortment is an evolutionary process common in viruses with segmented genomes. These viruses can swap whole genomic segments during cellular co-infection, giving rise to novel progeny formed from the mixture of parental segments. Since large-scale genome rearrangements have the potential to generate new phenotypes, reassortment is important to both evolutionary biology and public health research. However, statistical inference of the pattern of reassortment events from phylogenetic data is exceptionally difficult, potentially involving inference of general graphs in which individual segment trees are embedded. In this paper, we argue that, in general, the number and pattern of reassortment events are not identifiable from segment trees alone, even with theoretically ideal data. We call this fact the fundamental problem of reassortment, which we illustrate using the concept of the "first-infection tree," a potentially counterfactual genealogy that would have been observed in the segment trees had no reassortment occurred. Further, we illustrate four additional problems that can arise logically in the inference of reassortment events and show, using simulated data, that these problems are not rare and can potentially distort our observation of reassortment even in small data sets. Finally, we discuss how existing methods can be augmented or adapted to account for not only the fundamental problem of reassortment, but also the four additional situations that can complicate the inference of reassortment.
Asunto(s)
Genoma Viral , Filogenia , Virus Reordenados , Virus Reordenados/genética , Evolución Molecular , Modelos GenéticosRESUMEN
BACKGROUND: The efficacy of once-weekly (OW) glucagon-like peptide-1 receptor agonists (GLP-1RAs) has been established in several trials in people with type 2 diabetes mellitus (T2DM); however, real-world evidence on their effectiveness is limited. This study evaluated the effectiveness of OW GLP-1RA regarding glycemic and weight outcomes, and relative to DPP-4i in a comparator analysis. METHODS: This observational cohort study evaluated glycated hemoglobin (HbA1c) and weight outcomes in people with T2DM with two or more prescription claims for the same OW GLP-1RA using a pre-post study design (including for a semaglutide OW T2DM subgroup, hereafter referred to as semaglutide). Comparator analysis for the same outcome was performed for OW GLP-1RAs versus DPP-4i and semaglutide subgroup versus DPP-4i. A linked patient population from the IQVIA PharMetrics® Plus database and the Ambulatory Electronic Medical Records (AEMR) database was analyzed using data from January 2017 to April 2022. HbA1c and weight were assessed at baseline and at the end of the 12-month post-index period. Inverse probability of treatment weighting (IPTW) was used to adjust for imbalances in baseline patient characteristics in the comparator analysis. RESULTS: In the pre-post analysis, a greater numerical reduction in HbA1c and weight was observed for the semaglutide subgroup (N = 354) relative to the OW GLP-1RA cohort (N = 921). In the semaglutide subgroup, 52.5% and 34.2% of patients achieved HbA1c of < 7.0% and ≥ 5% weight loss, respectively. For the comparator analysis, the OW GLP-1RAs (N = 651) were significantly more effective (p < 0.001) in reducing HbA1c (- 1.5% vs. - 1.0%) and weight (- 3.2 kg vs. - 1.0 kg) than the DPP-4is (N = 431). Similarly, the semaglutide cohort (N = 251) also displayed more effectiveness (p < 0.001) in reducing HbA1c (- 1.7% vs. - 0.9%) and weight (- 4.1 kg vs. - 1.3 kg) than the respective DPP-4i cohort (N = 417). Patients initiating OW GLP-1RAs, including the semaglutide cohort, were at least twice as likely to achieve HbA1c and weight outcomes as well as composite outcomes compared with those initiating DPP-4is. CONCLUSION: The study reinforces that OW GLP-1RAs are more effective in glycemic control and weight reduction compared with DPP-4is in people with T2DM in the real-world setting. These findings align with the recommendation in the current guidelines for utilizing glucose-lowering treatment regimens that support weight-management goals in people with T2DM.
In type 2 diabetes mellitus (T2DM), glucagon-like peptide-1 receptor agonists (GLP-1RAs) are used for managing blood sugar levels and major adverse cardiovascular event risk reduction. In clinical trials, once-weekly (OW) GLP-1RAs showed better control of blood sugar levels and body weight than those administered daily, as well as another class of daily T2DM medications called dipeptidyl peptidase-4 inhibitors (DPP-4is). However, there is limited evidence of OW GLP-1RAs-based routine care to confirm these findings. This study gathered prescription and outcomes data for people with T2DM (January 2017April 2022) from two linked US databases. Body weight measurements and glycated hemoglobin (HbA1c) test results (measuring average blood sugar levels) were used to evaluate the effectiveness of OW GLP-1RAs (exenatide, dulaglutide, and semaglutide) via a pre-post analysis, and compare OW GLP-1RAs with DPP-4is. We found that treatment with semaglutide lowered body weight and blood sugar levels to a greater extent than OW GLP-1RAs in the pre-post analysis. In the comparator analysis, people receiving OW GLP-1RAs, including semaglutide, were at least twice as likely to achieve reduced HbA1c levels and body weight compared with those receiving DPP-4is. People receiving OW GLP-1RAs were three times more likely than those on DPP-4is to achieve the recommended target of HbA1c < 7.0% and weight loss ≥ 5%, while treatment with semaglutide increased this likelihood by > 4.6 times. This study shows clear benefits of OW GLP-1RAs, building on current evidence for integration of this treatment into overall management of T2DM.
Asunto(s)
Diabetes Mellitus Tipo 2 , Inhibidores de la Dipeptidil-Peptidasa IV , Humanos , Diabetes Mellitus Tipo 2/diagnóstico , Diabetes Mellitus Tipo 2/tratamiento farmacológico , Inhibidores de la Dipeptidil-Peptidasa IV/uso terapéutico , Hipoglucemiantes/uso terapéutico , Agonistas Receptor de Péptidos Similares al Glucagón , Control Glucémico , Pérdida de Peso , Dipeptidil-Peptidasas y Tripeptidil-Peptidasas , Receptor del Péptido 1 Similar al Glucagón/agonistasRESUMEN
For decades, mathematical models have been used to understand the course and outcome of malaria infections (i.e., infection dynamics) and the evolutionary dynamics of the parasites that cause them. A key conclusion of these models is that red blood cell (RBC) availability is a fundamental driver of infection dynamics and parasite trait evolution. The extent to which this conclusion holds will in part depend on model assumptions about the host-mediated processes that regulate RBC availability i.e., removal of uninfected RBCs and supply of RBCs. Diverse mathematical functions have been used to describe host-mediated RBC supply and clearance, but it remains unclear whether they adequately capture the dynamics of RBC supply and clearance during infection. Here, we use a unique dataset, comprising time-series measurements of erythrocyte (i.e., mature RBC) and reticulocyte (i.e., newly supplied RBC) densities during Plasmodium chabaudi malaria infection, and a quantitative data-transformation scheme to elucidate whether RBC dynamics conform to common model assumptions. We found that RBC clearance and supply are not well described by mathematical functions commonly used to model these processes. Furthermore, the temporal dynamics of both processes vary with parasite growth rate in a manner again not captured by existing models. Together, these finding suggest that new model formulations are required if we are to explain and ultimately predict the within-host population dynamics and evolution of malaria parasites.
RESUMEN
Reassortment is an evolutionary process common in viruses with segmented genomes. These viruses can swap whole genomic segments during cellular co-infection, giving rise to new viral variants. Large-scale genome rearrangements, such as reassortment, have the potential to quickly generate new phenotypes, making the understanding of viral reassortment important to both evolutionary biology and public health research. In this paper, we argue that reassortment cannot be reliably inferred from incongruities between segment phylogenies using the established remove-and-rejoin or coalescent approaches. We instead show that reassortment must be considered in the context of a broader population process that includes the dynamics of the infected hosts. Using illustrative examples and simulation we identify four types of evolutionary events that are difficult or impossible to reconstruct with incongruence-based methods. Further, we show that these specific situations are very common and will likely occur even in small samples. Finally, we argue that existing methods can be augmented or modified to account for all the problematic situations that we identify in this paper. Robust assessment of the role of reassortment in viral evolution is difficult, and we hope to provide conceptual clarity on some important methodological issues that can arise in the development of the next generation of tools for studying reassortment.
RESUMEN
Bagging (i.e., bootstrap aggregating) involves combining an ensemble of bootstrap estimators. We consider bagging for inference from noisy or incomplete measurements on a collection of interacting stochastic dynamic systems. Each system is called a unit, and each unit is associated with a spatial location. A motivating example arises in epidemiology, where each unit is a city: the majority of transmission occurs within a city, with smaller yet epidemiologically important interactions arising from disease transmission between cities. Monte Carlo filtering methods used for inference on nonlinear non-Gaussian systems can suffer from a curse of dimensionality as the number of units increases. We introduce bagged filter (BF) methodology which combines an ensemble of Monte Carlo filters, using spatiotemporally localized weights to select successful filters at each unit and time. We obtain conditions under which likelihood evaluation using a BF algorithm can beat a curse of dimensionality, and we demonstrate applicability even when these conditions do not hold. BF can out-perform an ensemble Kalman filter on a coupled population dynamics model describing infectious disease transmission. A block particle filter also performs well on this task, though the bagged filter respects smoothness and conservation laws that a block particle filter can violate.
RESUMEN
Over the past two decades, multiple countries with high vaccine coverage have experienced resurgent outbreaks of mumps. Worryingly, in these countries, a high proportion of cases have been among those who have completed the recommended vaccination schedule, raising alarm about the effectiveness of existing vaccines. Two putative mechanisms of vaccine failure have been proposed as driving observed trends: 1) gradual waning of vaccine-derived immunity (necessitating additional booster doses) and 2) the introduction of novel viral genotypes capable of evading vaccinal immunity. Focusing on the United States, we conduct statistical likelihood-based hypothesis testing using a mechanistic transmission model on age-structured epidemiological, demographic, and vaccine uptake time series data. We find that the data are most consistent with the waning hypothesis and estimate that 32.8% (32%, 33.5%) of individuals lose vaccine-derived immunity by age 18 y. Furthermore, we show using our transmission model how waning vaccine immunity reproduces qualitative and quantitatively consistent features of epidemiological data, namely 1) the shift in mumps incidence toward older individuals, 2) the recent recurrence of mumps outbreaks, and 3) the high proportion of mumps cases among previously vaccinated individuals.
Asunto(s)
Paperas , Vacunas , Humanos , Estados Unidos/epidemiología , Adolescente , Paperas/epidemiología , Paperas/prevención & control , Funciones de Verosimilitud , Virus de la Parotiditis/genética , Causalidad , Brotes de Enfermedades , VacunaciónRESUMEN
The spread of dengue and other arboviruses constitutes an expanding global health threat. The extensive heterogeneity in population distribution and potential complexity of movement in megacities of low and middle-income countries challenges predictive modeling, even as its importance to disease spread is clearer than ever. Using surveillance data at fine resolution from Rio de Janeiro, we document a scale-invariant pattern in the size of successive epidemics following DENV4 emergence. Using surveillance data at fine resolution following the emergence of the DENV4 dengue serotype in Rio de Janeiro, we document a pattern in the size of successive epidemics that is invariant to the scale of spatial aggregation. This pattern emerges from the combined effect of herd immunity and seasonal transmission, and is strongly driven by variation in population density at sub-kilometer scales. It is apparent only when the landscape is stratified by population density and not by spatial proximity as has been common practice. Models that exploit this emergent simplicity should afford improved predictions of the local size of successive epidemic waves.
Asunto(s)
Dengue , Epidemias , Brasil/epidemiología , Humanos , Densidad de Población , SerogrupoRESUMEN
We construct a family of genealogy-valued Markov processes that are induced by a continuous-time Markov population process. We derive exact expressions for the likelihood of a given genealogy conditional on the history of the underlying population process. These lead to a nonlinear filtering equation which can be used to design efficient Monte Carlo inference algorithms. We demonstrate these calculations with several examples. Existing full-information approaches for phylodynamic inference are special cases of the theory.
Asunto(s)
Algoritmos , Teorema de Bayes , Cadenas de Markov , Método de MontecarloRESUMEN
When modeling infectious diseases, it is common to assume that infection-derived immunity is either (1) non-existent or (2) perfect and lifelong. However there are many diseases in which infection-derived immunity is known to be present but imperfect. There are various ways in which infection-derived immunity can fail, which can ultimately impact the probability that an individual be reinfected by the same pathogen, as well as the long-run population-level prevalence of the pathogen. Here we discuss seven different models of imperfect infection-derived immunity, including waning, leaky and all-or-nothing immunity. For each model we derive the probability that an infected individual becomes reinfected during their lifetime, given that the system is at endemic equilibrium. This can be thought of as the impact that each of these infection-derived immunity failures have on reinfection. This measure is useful because it provides us with a way to compare different modes of failure of infection-derived immunity.
Asunto(s)
Probabilidad , PrevalenciaRESUMEN
Hosts defend themselves against pathogens by mounting an immune response. Fully understanding the immune response as a driver of host disease and pathogen evolution requires a quantitative account of its impact on parasite population dynamics. Here, we use a data-driven modeling approach to quantify the birth and death processes underlying the dynamics of infections of the rodent malaria parasite, Plasmodium chabaudi, and the red blood cells (RBCs) it targets. We decompose the immune response into 3 components, each with a distinct effect on parasite and RBC vital rates, and quantify the relative contribution of each component to host disease and parasite density. Our analysis suggests that these components are deployed in a coordinated fashion to realize distinct resource-directed defense strategies that complement the killing of parasitized cells. Early in the infection, the host deploys a strategy reminiscent of siege and scorched-earth tactics, in which it both destroys RBCs and restricts their supply. Late in the infection, a "juvenilization" strategy, in which turnover of RBCs is accelerated, allows the host to recover from anemia while holding parasite proliferation at bay. By quantifying the impact of immunity on both parasite fitness and host disease, we reveal that phenomena often interpreted as immunopathology may in fact be beneficial to the host. Finally, we show that, across mice, the components of the host response are consistently related to each other, even when infections take qualitatively different trajectories. This suggests the existence of simple rules that govern the immune system's deployment.
Asunto(s)
Interacciones Huésped-Parásitos/inmunología , Malaria/inmunología , Plasmodium chabaudi/patogenicidad , Reticulocitos/parasitología , Animales , Longevidad , Merozoítos/fisiología , Ratones , Modelos Teóricos , Plasmodium chabaudi/inmunología , Reticulocitos/inmunologíaRESUMEN
Importance: The United States has experienced a nationwide resurgence of pertussis since the mid-1970s, despite high estimated vaccine coverage. Short-lived immunity induced by diphtheria-tetanus-acellular pertussis (DTaP) vaccines in young children is widely believed to be responsible for this growing burden, but the duration of protection conferred by DTaP vaccines remains incompletely quantified. Objective: To assess the duration of immunity and the effectiveness of DTaP vaccines in US children. Design, Setting, and Participants: A mathematical, age-structured model of pertussis transmission, previously validated empirically on incidence data in Massachusetts, was used in this simulation study to assess the duration of DTaP immunity most consistent with the empirical values of the relative increase in the odds of acquiring pertussis from recent epidemiologic studies in the United States. The study included 5 simulated cohorts of children born between January 1, 2001, and December 31, 2005, followed up between the ages of 5 and 9 years (study period, January 1, 2006, to December 31, 2014). Statistical analysis was performed from May 1 to December 1, 2017. Interventions: Vaccination with DTaP according to the US immunization schedule, with a range of assumptions regarding the degree of waning immunity. Main Outcomes and Measures: Vaccine effectiveness and relative change in the odds of acquiring pertussis (odds ratio) in children aged 5 to 9 years, duration of DTaP immunity, and vaccine population-level impact. Results: This study found a marked association between the degree of waning immunity, vaccine effectiveness, and the odds ratio. Counterintuitively, the odds ratio was positively associated with vaccine effectiveness, as a consequence of nonlinear, age-assortative transmission dynamics. Based on the empirical odds ratios (1.33; 95% CI, 1.23-1.43), it was estimated that vaccine effectiveness exceeded 75% in children aged 5 to 9 years and that more than 65% of children remained immune to pertussis 5 years after the last DTaP dose. Conclusions and Relevance: The results of this study suggest that temporal trends in the odds of acquiring pertussis are an unreliable measure of the durability of vaccine-induced protection. They further demonstrate that DTaP vaccines confer imperfect, but long-lived protection. Control strategies should be based on the best available estimates of vaccine properties and the age structure of the transmission network.
Asunto(s)
Vacunas contra Difteria, Tétanos y Tos Ferina Acelular/inmunología , Inmunidad Innata , Inmunización Secundaria/métodos , Tos Ferina/prevención & control , Factores de Edad , Niño , Preescolar , Estudios de Seguimiento , Humanos , Esquemas de Inmunización , Incidencia , Masculino , Modelos Teóricos , Estudios Retrospectivos , Factores de Tiempo , Estados Unidos/epidemiología , Tos Ferina/epidemiología , Tos Ferina/transmisiónRESUMEN
Panel data, also known as longitudinal data, consist of a collection of time series. Each time series, which could itself be multivariate, comprises a sequence of measurements taken on a distinct unit. Mechanistic modeling involves writing down scientifically motivated equations describing the collection of dynamic systems giving rise to the observations on each unit. A defining characteristic of panel systems is that the dynamic interaction between units should be negligible. Panel models therefore consist of a collection of independent stochastic processes, generally linked through shared parameters while also having unit-specific parameters. To give the scientist flexibility in model specification, we are motivated to develop a framework for inference on panel data permitting the consideration of arbitrary nonlinear, partially observed panel models. We build on iterated filtering techniques that provide likelihood-based inference on nonlinear partially observed Markov process models for time series data. Our methodology depends on the latent Markov process only through simulation; this plug-and-play property ensures applicability to a large class of models. We demonstrate our methodology on a toy example and two epidemiological case studies. We address inferential and computational issues arising due to the combination of model complexity and dataset size. Supplementary materials for this article are available online.
RESUMEN
Inference using mathematical models of infectious disease dynamics can be an invaluable tool for the interpretation and analysis of epidemiological data. However, researchers wishing to use this tool are faced with a choice of models and model types, simulation methods, inference methods and software packages. Given the multitude of options, it can be challenging to decide on the best approach. Here, we delineate the choices and trade-offs involved in deciding on an approach for inference, and discuss aspects that might inform this decision. We provide examples of inference with a dataset of influenza cases using the R packages pomp and rbi.
RESUMEN
We present new evidence that the immunity conferred against pertussis by the DTaP acellular vaccine wanes more slowly than widely believed.
Asunto(s)
Vacunas contra Difteria, Tétanos y Tos Ferina Acelular , Tos Ferina , Anticuerpos Antibacterianos , Humanos , Cobertura de VacunaciónRESUMEN
The resurgence of pertussis over the past decades has resulted in incidence levels not witnessed in the United States since the 1950s. The underlying causes have been the subject of much speculation, with particular attention paid to the shortcomings of the latest generation of vaccines. We formulated transmission models comprising competing hypotheses regarding vaccine failure and challenged them to explain 16 years of highly resolved incidence data from Massachusetts, United States. Our results suggest that the resurgence of pertussis is a predictable consequence of incomplete historical coverage with an imperfect vaccine that confers slowly waning immunity. We found evidence that the vaccine itself is effective at reducing overall transmission, yet that routine vaccination alone would be insufficient for elimination of the disease. Our results indicated that the core transmission group is schoolchildren. Therefore, efforts aimed at curtailing transmission in the population at large, and especially in vulnerable infants, are more likely to succeed if targeted at schoolchildren, rather than adults.
Asunto(s)
Vacuna contra la Tos Ferina/uso terapéutico , Humanos , Vacuna contra la Tos Ferina/inmunología , Estados Unidos , Vacunación/estadística & datos numéricos , Cobertura de Vacunación/estadística & datos numéricosRESUMEN
The recent resurgence of pertussis in England and Wales has been marked by infant deaths and rising cases in teens and adults. To understand which age cohorts are most responsible for these trends, we employed three separate statistical methods to analyze high-resolution pertussis reports from 1982 to 2012. The fine-grained nature of the time-series allowed us to describe the changes in age-specific incidence and contrast the transmission dynamics in the 1980s and during the resurgence era. Our results identified infants and school children younger than 10â¯years of age as a core group, prior to 2002: pertussis incidence in these populations was predictive of incidence in other age groups. After 2002, no core groups were identifiable. This conclusion is independent of methodology used. Because it is unlikely that the underlying contact patterns substantially changed over the study period, changes in predictability likely result from the introduction of more stringent diagnostics tests that may have inadvertently played a role in masking the relative contributions of core transmission groups.