Academic literature on the topic 'Disease modelling, Systems analysis, epidemiology, Model evaluation'

Introduction. Outbreaks of infectious diseases and the COVID-19 pandemic in particular pose a serious public health challenge. The other side of the challenge is always opportunity, and today such opportunities are information technology, decision making systems, best practices of proactive management and control based on modern methods of data analysis (data driven decision making) and modeling. The article reviews the prospects for the use of publicly available software in modeling epidemiological trends. Strengths and weaknesses, main characteristics and possible aspects of application are considered. The purpose of the article is to review publicly available health software. Give situations in which one or another approach will be useful. Segment and determine the effectiveness of the underlying models. Note the prospects of high-performance computing to model the spread of epidemics. Results. Although deterministic models are ready for practical use without specific additional settings, they lose comparing to other groups in terms of their functionality. To obtain evaluation results from stochastic and agentoriented models, you first need to specify the epidemic model, which requires deeper knowledge in the field of epidemiology, a good understanding of the statistical basis and the basic assumptions on which the model is based. Among the considered software, EMOD (Epidemiological MODelling software) from the Institute of Disease Modeling is a leader in functionality. Conclusions. There is a free access to a relatively wide set of software, which was originally developed by antiepidemiological institutions for internal use in decision-making, however was later opened to the public. In general, these programs have been adapted to increase their practical application. Got narrowed focus on potential issues. The possibility of adaptive use was provided. We can note the sufficient informativeness and convenience of using the software of the group of deterministic methods. Also, such models have a rather narrow functional focus. Stochastic models provide more functionality, but lose some of their ease of use. We have the maximum functionality from agentoriented models, although for their most effective use you need to have the appropriate skills to write program code. Keywords: epidemiological software, deterministic modeling, stochastic modeling, agentoriented mode-ling, high performance computing, decision making systems.

Pertussis is an infectious respiratory disease for which mass vaccination is an effective preventive strategy. In many developed countries, where high vaccination coverage has been maintained for approximately 50 years, re-emergence of the disease has been observed in all age groups. In the municipality of Rio de Janeiro (RJ), where vaccination started in the 1980s, surveillance data show no sign of disease re-emergence. We developed a mathematical model that incorporates the major demographic aspects of a large urban centre in a developing nation, in addition to the most important epidemiological aspects of disease transmission. Parameter values were estimated based on RJ demographic and vaccine coverage data. Overall, all vaccination strategies determined a major decrease (over 95% decrease when compared to the pre-vaccine era) in the incidence of primary infections (occurring in individuals who have never been immunized through infection or vaccine). On the other hand, the strategies (a) three doses at age 2–11 months, (b) three doses plus booster at age 12–23 months, (c) three doses plus booster at age 4–5 years, and (d) three doses plus both boosters, differently affected the incidence of secondary infections (occurring in previously infected/vaccinated individuals). Given that the immunity against pertussis wanes with time and that the infectious agent has not been eliminated from the population, it is expected that pertussis will continue to be a problem in RJ. Actually, since immunity acquired from vaccine wanes faster than disease-acquired immunity and the possibility of natural boosters has decreased with mass vaccination, an increase in the incidence of secondary infections among older age groups is expected (and predicted by the model). Possible explanations as to why this dynamics is not captured by the RJ surveillance system are discussed. A poorly effective surveillance system and a lack of awareness regarding loss of immunity and the possibility of pertussis infection in older age groups are among them. Finally, we bring attention to the need of (i) field studies for the measurement of pertussis incidence in adolescents and adults; (ii) better understanding of the transmission dynamics currently occurring in RJ, and (iii) re-evaluation of vaccination strategies with the possible introduction of acellular vaccines for the vaccination of older individuals.

A cardinal challenge in epidemiological and ecological modelling is to develop effective and easily deployed tools for model assessment. The availability of such methods would greatly improve understanding, prediction and management of disease and ecosystems. Conventional Bayesian model assessment tools such as Bayes factors and the deviance information criterion (DIC) are natural candidates but suffer from important limitations because of their sensitivity and complexity. Posterior predictive checks, which use summary statistics of the observed process simulated from competing models, can provide a measure of model fit but appropriate statistics can be difficult to identify. Here, we develop a novel approach for diagnosing mis-specifications of a general spatio-temporal transmission model by embedding classical ideas within a Bayesian analysis. Specifically, by proposing suitably designed non-centred parametrization schemes, we construct latent residuals whose sampling properties are known given the model specification and which can be used to measure overall fit and to elicit evidence of the nature of mis-specifications of spatial and temporal processes included in the model. This model assessment approach can readily be implemented as an addendum to standard estimation algorithms for sampling from the posterior distributions, for example Markov chain Monte Carlo. The proposed methodology is first tested using simulated data and subsequently applied to data describing the spread of Heracleum mantegazzianum (giant hogweed) across Great Britain over a 30-year period. The proposed methods are compared with alternative techniques including posterior predictive checking and the DIC. Results show that the proposed diagnostic tools are effective in assessing competing stochastic spatio-temporal transmission models and may offer improvements in power to detect model mis-specifications. Moreover, the latent-residual framework introduced here extends readily to a broad range of ecological and epidemiological models.

Abstract Purpose Three-dimensional (3D) gallbladder (GB) geometrical models are essential to GB motor function evaluation and GB wall biomechanical property identification by employing finite element analysis (FEA) in GB disease diagnosis with ultrasound systems. Methods for establishing such 3D geometrical models based on static two-dimensional (2D) ultrasound images scanned along the long-axis/sagittal and short-axis/transverse cross-sections in routine GB disease diagnosis at the beginning of emptying phase have not been documented in the literature so far. Methods Based on two custom MATLAB codes composed, two images were segmented manually to secure two sets of the scattered points for the long- and short-axis GB cross-section edges; and the points were best fitted with a piecewise cubic spline function, and the short-axis cross-section edges were lofted along the long-axis to yield a 3D geometrical model, then GB volume of the model was figured out. The model was read into SolidWorks for real surface generation and involved in ABAQUS for FEA. Results 3D geometrical models of seven typical GB samples were established. Their GB volumes are with 15.5% and − 4.4% mean errors in comparison with those estimated with the ellipsoid model and sum-of-cylinders method but can be correlated to the latter very well. The maximum first principal in-plane stress in the 3D models is higher than in the ellipsoid model by a factor of 1.76. Conclusions A numerical method was put forward here to create 3D GB geometrical models and can be applied to GB disease diagnosis and GB shape analysis with principal component method potentially in the future.

Abstract Background Antimicrobial resistance (AMR) is a global threat to effective prevention and treatment of an ever-increasing range of infections. Pneumococcal conjugate vaccines (PCV) used in infant national immunization programs have been shown to decrease AMR pneumococci. Cost-effectiveness models evaluating the value for money of PCV programs have not considered the economic impact of reducing antimicrobial prescribing or prolonged infections due to treatment failures. Standardized frameworks are needed for models to address outcomes and impact on health resource utilization related to AMR. Methods We developed a conceptual modeling methodology suitable for a health economic evaluation of an infant PCV program. We considered impact of PCVs on pneumococcal disease (PD) specifically related to clinical management of AMR-PD, including AMR epidemiology, antibiotic prescribing patterns, and healthcare resource utilization. Model inputs were evaluated regarding optimal and available data sources considering the complex nature of AMR at the national, regional, and global level. Results The proposed framework considers impact of PCVs on antimicrobial prescribing due to invasive pneumococcal disease (IPD), community acquired pneumonia (CAP), and acute otitis media (AOM) across 3 pathways (Figure 1). The population and pathogen-level pathway describe epidemiology and vaccine impact. The care level pathway describes clinical disease management. The health outcomes pathway characterizes resistant or successfully treated PD costs and quality of life. Conceptual Economic Model Methodology Conclusion We present a generalizable methodology to quantify impact of PCVs on cases and outcomes of PD related to AMR. Modelling vaccine-preventable burden of AMR-PD requires data extrapolations and assumptions due to the myriad of interconnected pathways (i.e. microbiology, epidemiology, environment, health systems). Further work is needed to validate assumptions and linkages across incomplete data sources. Disclosures Raymond Farkouh, PhD, Pfizer (Employee) Arianna Nevo, MPH, Pfizer, Inc. (Other Financial or Material Support, I am an employee of IQVIA. IQVIA received funding from Pfizer to carry out the project.) Jennifer Uyei, PhD, MPH, Pfizer, Inc. (Other Financial or Material Support, I am an employee of IQVIA. IQVIA received funding from Pfizer to carry out the project.) Cassandra Hall-Murray, PharmD, Pfizer, Inc. (Employee) Joseph Lewnard, PhD, Pfizer, Inc. (Consultant, Grant/Research Support, Advisor or Review Panel member) Matthew Wasserman, MSc., Pfizer Inc. (Employee)

The safe removal of disease-carrying human waste is the objective of all sanitation systems and the limiting of air pressure transients within the system remains a significant part of current codes and regulations. The water trap seal offers fundamental protection and is the system’s sole barrier between the public sewer network and habitable space inside a building. Modelling water trap seal responses to air pressure fluctuations offers an opportunity to analyse whole system performance, but the quality of the data depends on the accuracy of the modelling technique and that of the defining inputs. AIRNET, a 1D Method of Characteristics based model, enables rapid whole system testing; however, the present boundary condition for the water trap seal within the model is based solely on steady state conditions, ignoring system dynamics. Computational fluid dynamics offers an opportunity to numerically evaluate the flow patterns within the trap seal in response to applied air pressure transients. This research confirms the importance of the rate of rise, and hence frequency of air pressure transients incident on water trap seals and relates this to potential vulnerabilities of different device geometries, particularly the ratio between inner and outer wall length. The research led to the development of a dynamic velocity decrement model encapsulating unsteady friction and separation losses linked to device geometry for the first time. The development of a frequency-dependent internal energy term Δ v, suitable for inclusion in AIRNET provides the capability to predict more realistic water trap response to air pressure transients over a range of air pressure transient frequencies likely to cause problems: 1 Hz to 8 Hz. Practical application: Whole system modelling can greatly improve the ability of design engineers to fully simulate the operation of a building drainage system in a realistic way. The work described in this paper improves the accuracy of whole system models by evaluating water dynamic responses to air pressure transients using a range of techniques including computational fluid dynamics and more traditional 1D finite difference method of characteristics models. The work also paves the way for more robust evaluation of building drainage products through in-depth investigation of the fluid mechanics associated with their operation.

The widespread, and in many countries unprecedented, use of non-pharmaceutical interventions (NPIs) during the COVID-19 pandemic has highlighted the need for mathematical models which can estimate the impact of these measures while accounting for the highly heterogeneous risk profile of COVID-19. Models accounting either for age structure or the household structure necessary to explicitly model many NPIs are commonly used in infectious disease modelling, but models incorporating both levels of structure present substantial computational and mathematical challenges due to their high dimensionality. Here we present a modelling framework for the spread of an epidemic that includes explicit representation of age structure and household structure. Our model is formulated in terms of tractable systems of ordinary differential equations for which we provide an open-source Python implementation. Such tractability leads to significant benefits for model calibration, exhaustive evaluation of possible parameter values, and interpretability of results. We demonstrate the flexibility of our model through four policy case studies, where we quantify the likely benefits of the following measures which were either considered or implemented in the UK during the current COVID-19 pandemic: control of within- and between-household mixing through NPIs; formation of support bubbles during lockdown periods; out-of-household isolation (OOHI); and temporary relaxation of NPIs during holiday periods. Our ordinary differential equation formulation and associated analysis demonstrate that multiple dimensions of risk stratification and social structure can be incorporated into infectious disease models without sacrificing mathematical tractability. This model and its software implementation expand the range of tools available to infectious disease policy analysts.

4596 Background: Small bowel adenocarcinoma is a rare malignancy and is often associated with poor outcome. The impact that the number of positive and negative lymph nodes (LN) have upon survival following curative resection has not been studied. Methods: Patients aged 18–90 with adenocarcinoma of the small intestine diagnosed between 1988 and 2005 were identified from SEER data (ver. 2008). Disease-specific survival (DSS) outcomes were determined through 12/2005. Cox proportional hazards regression analyses were performed after adjusting for age, sex, race, T stage, grade, and primary site. Stage I-II cases were categorized by total LN examined (1–8, 9–12, and >12). Stage III cases were evaluated using cut-point analysis to determine the number of positive LN that predicted outcomes. This result was then compared to the predictive value of the ratio of positive to total LN (LNR) using the chi-square statistic. Results: 1,991 patients were identified in the SEER database. Survival among stage I/II patients (n=1,216) was dependent upon the total number of LN assessed. 5-year DSS for stage II patients was 66%, 82% (HR 0.52 95% CI .33-.84), and 88% (HR 0.38, 95% CI .23-.61) for 1–8, 9–12, >12 LN, respectively. The optimal cutpoint of positive LN for stage III disease (n=775) was <3 compard to ≥3 with 5 year DSS of 58% vs. 37% (HR 1.49, 95% CI 1.15–1.92, P=0.002), respectively. Among stage III patients, the LNR was even more predictive of survival than stratification by the number of positive lymph nodes as demonstrated by an improved chi-square statistic for the multivariate model (78.8 vs 63.1, P=0.0005). Conclusions: As noted in colon cancer, the total number of LN assessed has considerable influence upon survival in stage I, II and III small bowel adenocarcinoma. Stratifying stage III small bowel adenocarcinoma into those with <3 and ≥3 positive lymph nodes significantly improves prognostication for these patients and future staging systems should incorporate the number of positive nodes into nodal staging. The use of LNR may provide additional prognostic information. No significant financial relationships to disclose.

Objectives Sub-Saharan Africa faces a shortage of skilled epidemiologists to prevent, detect, and respond to health threats. Tanzania has implemented one of the first Centers for Disease Control and Prevention Field Epidemiology Training Program (FETP) Intermediate courses in Africa. This course aims to strengthen health workforce capacity in surveillance system assessment, outbreak investigation, and evaluation, prioritizing HIV control. We conducted an outcome evaluation of this new course. Methods We used a pre/post evaluation design using data from 4 cohorts of trainees who took the FETP Intermediate course from 2017 to 2020. We conducted knowledge assessments before and after each cohort and combined those results. Outcomes included knowledge and self-rated competency and trends in integrated disease surveillance and response (IDSR) data. We collected data through tests, field assignments, exit interviews, and data audits. We compared the mean change in pre-/posttest scores using linear regression and 95% CIs. We used content analysis to summarize exit interviews. Results Fifty-three FETP trainees from 10 regions enrolled in the FETP Intermediate course, and 52 (99.0%) completed the course. We found substantial increases in mean knowledge (44.0 to 68.0 points) and self-rated competency (4.14 to 4.43) scores before and after the course. Trainees evaluated 52 surveillance systems and 52 district HIV care programs, and 39 (75.0%) trainees participated in outbreak investigations. From before to after cohort 1, timeliness and completeness of IDSR reports increased from 4.2% to 52.1% and from 27.4% to 76.5%, respectively. Course strengths were quality of instruction, individualized mentoring, and practical skills gained. Challenges were mentor availability, limited time for data analysis practice, and balancing work and field assignments. Conclusions The Tanzania FETP Intermediate course substantially improved trainee knowledge and helped to improve local data quality and reporting. This course is a promising model to strengthen subnational capacity to prevent, detect, and respond to public health threats in Africa.

Bluetongue (BT) is among the World Animal Health Organization (OIE) listed diseases due to its potential for rapid spread and serious economic impact on livestock. Because of its epidemiology, in Europe, only Southern countries were affected by the disease in the past. However in the latter half of 2006, an unprecedented outbreak of bluetongue virus (BTV) serotype 8 occurred in North-Western European countries. To define potential regions that are at risk for BT epidemics it is essential to study vector distribution and abundance. This study focused on BT vector spread, mostly in Austria. The objective was to produce risk maps with the more likely areas for vector occurrence and thus to support BT prevention and control. The introductory review gives an overview of the epidemiology of the disease with a focus on the vectors, the recent outbreaks in North-Western Europe, and the importance of statistical model­ling and geographical information systems (GIS) in predicting, preventing and controlling BT. The statistical analysis was mainly based on data from the Austrian entomological surveillance system, weather stations and topo­graphical information. A multiple linear regression model was fitted to the data to predict the occurrence of BTV vectors and subsequently to create risk maps for the whole country. Despite the fact that the limited nature of the data does not allow precise estimation, in general the models indicated that vectors occurred in preferential areas where they could be very abundant. A more detailed analysis should be carried out with a multidisciplinary team including epidemiologists, biologists, meteorologists, ento­mologists, and statisticians, so that the complexity of BT epide­miology may be better understood, and a more efficient process of prevention and control of the disease may be set up.

L’antracnosi della vite (causato da Elsinoe ampelina), il mariciume dell’uva (o ripe rot causato da Colletotrichum spp.) e la carie bianca (o white rot causato da Coniella diplodiella) sono malattie che minacciano seriamente la produzione viticola in Cina. Infatti, il loro controllo implica la ripetuta applicazione di fungicidi durante la stagione produttiva. Nella presente tesi, le conoscenze disponibili sull’antracnosi e il marciume dell’uva (ripe rot) sono state raccolte dalla letteratura, analizzate e sintetizzate attraverso l’analisi dei sistemi, per sviluppare modelli meccanicistici in cui le variabili ambientali guidano le predizioni del modello. A differenza di quanto ottenuto per questi patogeni, la revisione della letteratura per la carie bianca (white rot) ha evidenziato una carenza di informazioni per alcuni importanti aspetti epidemiologici di questa malattia. Pertanto, sono stati eseguiti diversi esperimenti in laboratorio per studiare l’effetto di diverse condizioni ambienti sull’infezione degli acini, il periodo di incubazione, il periodo di latenza e la dinamica di sporulazione. Sulla base dei risultati di questi esperimenti, è stato quindi sviluppato un modello meccanicistico per la carie bianca che considera l’intero ciclo vitale del patogeno. I tre modelli meccanicistici sviluppati sono stati validati confrontando le predizioni con dati reali indipendenti; i modelli hanno mostrato una buona capacità e accuratezza nella rappresentazione delle epidemie in condizioni reali. I modelli sviluppati in questa tesi forniscono la base per una migliore protezione del vigneto da queste malattie.
Grapevine anthracnose (caused by Elsinoe ampelina), ripe rot (caused by Colletotrichum spp.) and white rot (caused by Coniella diplodiella) are serious threats in many vineyards of China, and their controls require repeated application of fungicides. In the present dissertation, the available knowledge on grapevine anthracnose and ripe rot were retrieved from literature, analyzed, and synthesized to develop weather-driven, mechanistic models for indicated two diseases based on system analysis. Unlike the previous two diseases, the literature review provided incomplete information about some important aspects of epidemiology of grape white rot. Therefore, several artificial inoculation experiments were conducted to investigate the effects of environmental factors on epidemiological parameters of white rot, including berry infection, incubation, latency and sporulation dynamics. Subsequently, a process-based mechanistic model that accounts for the entire life cycle of pathogen was constructed by using these experimental data. The above three mechanistic models were validated by comparing with independent datasets, and provided good ability and accuracy to represent the real epidemiological systems. The models developed in this dissertation provide a basis for better scheduling crop protection actions in vineyards.

In patologia vegetale, un modello epidemiologico è una rappresentazione semplificata delle relazioni tra patogeno, ospite e ambiente che determinano se, quando e come un’epidemia si sviluppa nel tempo. La presente tesi raccoglie lo sviluppo di modelli meccanicistici, dinamici, guidati da variabili ambientali (weather-driven), che possono essere utilizzati per la difesa di precisione dalle principali malattie di frumento e leguminose in rotazione colture. Tramite l’applicazione dell’analisi dei sistemi alle informazioni contenute nella letteratura scientifica, ogni patosistema è stato analizzato e concettualizzato per lo sviluppo teoretico e matematico del modello. I seguenti patogeni sono stati considerati per lo sviluppo dei modelli: i) Ascochyta rabiei agente causale della rabbia (Ascochyta blight) del cece; ii) Puccinia graminis f.sp. tritici agente causale della ruggine nera del grano; iii) Sclerotinia sclerotiorum, una specie polifaga e agente causale di diverse malattie di leguminose e colture industriali (es. marciume bianco di fagiolo e soia, muffa bianca del colza, sclerotiniosi del girasole). La capacità previsionale dei modelli è stata valutata utilizzando dati reali indipendenti raccolti in diversi condizioni ambientali ed epidemiologiche. Dal confronto tra le predizioni e i dati reali di campo, i modelli si sono rivelati accurati e robusti e, quindi, utilizzabili dagli agricoltori nei processi decisionali per la protezione delle colture. La presente tesi contiene anche i risultati di una revisione della letteratura sulle esigenze termiche del genere Colletotrichum, contenente numerosi agenti causali dell’antracnosi di leguminose e diverse colture industriali. L’effetto della temperatura su quattro processi biologici (crescita miceliale, germinazione delle spore, infezione e sporulazione) è stato modellizzato per i maggiori gruppi filogenetici di Colletotrichum spp.. Questo lavoro getta le basi per lo sviluppo di un modello meccanicistico, dinamico, weather-driven generalizzato per Colletotrichum spp. sulla base delle similarità entro i gruppi filogenetici.
A plant disease model is a simplification of the relationships between pathogen, host, and environment that determine whether and how an epidemic develops over time. The present dissertation aims to develop mechanistic, dynamic, weather-driven models, which are suitable to be applied in precision crop protection, for important diseases affecting wheat and legumes in a crop rotation scenario. By exploitation of literature and application of system analysis, information concerning the pathosystem were acquired and analyzed to conceptualize and develop the model both theoretically and mathematically. The following pathogens were considered: i) Ascochyta rabiei causing Ascochyta blight in chickpea; ii) Puccinia graminis f.sp. tritici causing stem (or black) rust of wheat; iii) Sclerotinia sclerotiorum, a polyphagous specie causing diseases in several legumes and industrial crops (e.g., white mold of white bean and soybean, stem rot of canola, head rot of sunflower). Models were evaluated using independent data for their ability to predict the occurrence and development of epidemics, under different environmental conditions. The comparison of model predictions versus real data observed in fields showed that models could be considered accurate and robust and, therefore, they may be used to help growers in making decisions to efficiently protect their crops. The present dissertation contains also results of a literature review carried out on temperature requirements of Colletotrichum spp., which involves several species causing anthracnose on legumes and several industrial crops. Temperature-dependent equations were developed for four biological processes (mycelial growth, germination of spores, spore infection, and spore production) of major phylogenetic clades of Colletotrichum spp.. This work may lay the foundation for the development of a general, mechanistic, dynamic, weather-driven model for Colletotrichum spp. based on the intra-clade similarities.