Mapping of mortality rates has been a valuable public health tool. We describe novel Bayesian methods for constructing maps which do not depend on a post stratification of the estimated rates. We also construct posterior modal maps rather than posterior mean maps. Our methods are illustrated using mortality data from chronic obstructive pulmonary diseases (COPD) in the continental United States. Poisson regression models have attracted much attention in the scientific community for their superiority in modeling rare events (including mortality counts from COPD). Christiansen and Morris (JASA 1997) described a hierarchical Bayesian model for heterogeneous Poisson counts under the exchangeability assumption. We extend this model to include latent classes (groups of similar Poisson rates unknown to an investigator). Also, it is standard practice to construct maps using quantiles (e.g., quintiles) of the estimated mortality rates. For example, based on quintiles, the mortality rates are cut into 5 equal size groups, each containing $20\%$ of the data, and a different color is applied to each of them on the map. A potential problem is that, this method assumes an equal number of data in each group, but this is often not the case. The latent class model produces a method to construct maps without using quantiles, providing a more natural representation of the colors. Typically, for rare events, the posterior densities of the rates are skewed, making the posterior mean map inappropriate and inaccurate. Thus, although it is standard practice to present the posterior mean maps, we also develop a method to provide the joint posterior modal map (i.e., the map with the highest posterior probability over the ensemble). For the COPD data, collected 1988-1992 over 798 health service areas, we use Markov chain Monte Carlo methods to fit the model, and an output analysis is used to construct the new maps.

Programinės įrangos kūrimo procese reikalavimų rinkimo ir jų analizės etapai yra patys svarbiausi, nuo jų priklauso tolesnių programinės įrangos kūrimo etapų įgyvendinimas. Informacinių sistemų specifikavimui dažniausiai naudojamos projektavimo kalbos (UML, Z, DFD, OCL ir t.t.), kurių supratimui reikalingos specifinės žinios. Užsakovo projektavimo kalbos žinios dažnai yra minimalios, tad jis pasitiki projektuotojo žiniomis ir pateiktų reikalavimų interpretavimu. Projektavimo etape tinkamai užsakovo nesuprastų, formalia kalba specifikuotų, sistemos projektinių modelių netikslumai, dažnai išryškėja tik realizavimo etape. Klaidingai IT specialisto interpretuotų žinių padarinių taisymas yra brangus ir pailgina programinio produkto kūrimo trukmę. Šiam nesusišnekimui išvengti buvo pasiūlyta dalykinę sritį aprašinėti veiklos taisyklėmis natūralios kalbos šablonais. Veiklos taisyklės yra struktūrizuotos ir vienareikšmiškai interpretuojamos tiek IT specialistų, tiek verslo atstovų. Šis dalykinės srities aprašymo metodas yra labai neseniai pasirodęs ir dar nėra nusistovėjusių standartų jo taikyme. Šiame darbe bus kuriamas algoritmas, kuris veiklos taisyklių pagrindu patikslins nuo skaičiavimų nepriklausomo (arba nuo platformos nepriklausomo) lygmens duomenų modelius. Tuo savo ruoštu paspartindamas projektinių modelių kūrimą ir patikslinimą automatizuojant veiklos taisyklių ir klasių modelio analizę. / Software development process of the collection and analysis phases are the key, since they belong to the software development phases of implementation. Information systems are the most used in the design specification language (UML, Z, DFD, OCL, etc.), understanding the need for specific knowledge. Customer's design language is often minimal, so he is confident in the knowledge the designer and the requirements of the interpretation. The design phase of the customer due misunderstood formally specified language, system design models, inaccuracies, often become apparent only in the realization phase. Erroneous interpretation of the IT professional knowledge effects editing is expensive and prolonged the duration of software product development. This misunderstood was proposed to avoid the subject area specificate operational rules of natural-language ralios pattern. Activities are structured and unambiguous interpretation of the IT professionals and business representatives. The Subject field of the description method is a very recent and has not established standards for its application. This work will be developed in the algorithm, the business rules-based adjustments to the calculations of independent (or platform-independent) level of data models. Meanwhile, prepare their own models for accelerating the development of design automation and to verify rules and model of class analysis.

No description available.

Rift Valley fever (RVF) is a mosquito-borne zoonotic disease caused by a virus of the family Bunyaviridae, genus Phlebovirus. It is responsible for extensive outbreaks of disease in livestock in Africa with significant mortality and economic impact. Virus neutralization is considered the gold standard for confirming Rift Valley fever virus (RVFV) infection but the procedure is time consuming and expensive. Real-time reverse transcription-polymerase chain reaction (rRT-PCR), histopathology, and immunohistochemistry (IHC) are the diagnostic methods most often used in South Africa to confirm or exclude a diagnosis of RVF in necropsied animals. Validated estimates of diagnostic accuracy of these tests, in naturally infected livestock, however, have not been published. The objective of this study was to estimate the diagnostic sensitivity and specificity of rRT-PCR, histopathology, and IHC using Bayesian latent class methods in the absence of a gold standard. A secondary objective was to estimate stratum-specific values based on species, age, degree of specimen autolysis, and the presence/absence of tissue pigments. The Sensitivity (Se) and Specificity (Sp) of qRT-PCR were 97.4% (95% credibility interval (CI): 95.2% - 98.8%) and 71.7% (95% CI: 65% - 77.9%) respectively. The extraordinary analytical sensitivity of PCR makes this test very susceptible to false positive reactions, and thus reduced specificity. This is more likely during large-scale epidemics due to crosscontamination of specimens at necropsy facilities or testing laboratories. The Se and Sp of histopathology were 94.6% (95% CI: 91% - 97.2%) and 92.3% (95% CI: 87.6% - 95.8%) respectively. Single cases of RVF could be confused with acute poisoning with plants, bacterial septicaemias, and viral diseases such as infectious bovine rhinotracheitis and Wesselsbron disease. Most of these conditions, however, can be excluded using histological examination of the liver, special stains, bacterial culture, and toxicological or serological investigations. The Se and Sp of IHC were 97.6% (95% CI: 93.9% - 99.8%) and 99.4% (95% CI: 96.9% - 100%) respectively. Immunohistochemistry is highly specific because characteristic positive immunolabelling of the cytoplasm of hepatocytes can be correlated with the presence of hepatocellular injury typical for RVFV infection. False negative results are sometimes obtained with IHC because of reader error or loss of the antigenic epitopes due to advanced autolysis. Scant positive immunolabelling might be missed or viral proteins might be absent from sections of liver with advanced hepatocellular damage. The stratified analysis suggested differences in test accuracy in foetuses and severely autolysed specimens. The Sp of histopathology in foetuses (83.0%) was 9.3% lower than the value obtained for the sample population (92.3%). Lesions in some foetuses are more subtle and the typical eosinophilic intranuclear inclusions are often difficult to detect. In severely autolysed specimens, the Se of IHC decreased by 16.1% and the Sp of rRT-PCR by 17.4%. There is no plausible biological explanation for this decrease in the Sp of rRTPCR since the RNA of RVFV is resistant to degradation in autolysed tissues. Conversely, the antibody used to detect RVFV using IHC detects epitopes raised against nucleoproteins of the virus and it is possible that viral proteins become too widely dispersed and/or degraded in autolysed tissues to detect by light microscopy. It is possible that the marked decrease in Se of histopathology and IHC in severely autolysed specimens caused an apparent decrease in Sp of rRT-PCR, due to the latent class method. In conclusion, the high estimated Sp (99.4%) of IHC and the low Sp of rRT-PCR (71.3%) suggests that the definitive diagnosis or exclusion of RVF should not rely on a single PCR test and that IHC would be an effective confirmatory test for rRT-PCR positive field cases necropsied during an epidemic. Immunohistochemistry results from severely autolysed specimens, however, should be interpreted with caution and aborted foetuses in areas endemic for RVF should be screened using a variety of tests. The diagnostic Se and Sp of histopathology was much higher than expected confirming the value of routine post mortem examinations and histopathology of liver specimens. The most feasible RVF testing option in areas that do not have suitably equipped PCR laboratories, and where disease is often not detected in livestock until after human cases have been diagnosed, would be routine histopathology screening with IHC confirmation. Key Words: Rift Valley fever; Rift Valley fever virus; Bayesian; latent-class model; real-time reverse transcription-polymerase chain reaction; immunohistochemistry; histopathology; diagnosis; sensitivity; specificity. / Dissertation (MSc)--University of Pretoria, 2014. / gm2014 / Paraclinical Sciences / unrestricted

Title: Joint Models for Longitudinal and Time-to-Event Data Author: Jana Vorlíčková Department: Department of Probability and Mathematical Statistics Supervisor: doc. RNDr. Arnošt Komárek, Ph.D., Department of Probability and Mathematical Statistics Abstract: The joint model of longitudinal data and time-to-event data creates a framework to analyze longitudinal and survival outcomes simultaneously. A commonly used approach is an interconnection of the linear mixed effects model and the Cox model through a latent variable. Two special examples of this model are presented, namely, a joint model with shared random effects and a joint latent class model. In the thesis we focus on the joint latent class model. This model assumes an existence of latent classes in the population that we are not able to observe. Consequently, it is assumed that the longitudinal part and the survival part of the model are independent within one class. The main intention of this work is to transfer the model to the Bayesian framework and to discuss an estimation procedure of parameters using a Bayesian statistic. It consists of a definition of the model in the Bayesian framework, a discussion of prior distributions and the derivation of the full conditional distributions for all parameters of the model. The model's ability to...

<p>Commercial and regional brewers are increasingly investing in sustainability equipment that reduces input use, operating costs, and environmental impact. These technologies often require significant upfront costs that can limit market access to microbreweries. One potential solution for these brewers is to market their product as sustainable and charge a premium for their product to offset some of the costs. A stated preference choice experiment of a nationally-representative sample is undertaken to elicit consumer willingness-to-pay (WTP) for sustainability attributes in beer, thus determining whether a market for sustainably-made beer exists. The facets of sustainability, including water reduction, energy reduction, and landfill diversion, are portrayed through eco-labels affixed the front of the primary packaging (aluminum can or glass bottle). Multiple specifications are employed to handle model shortcomings and incorporate discrete heterogeneity. Across all model specifications, <a>consumers show a positive and statistically significant marginal WTP for landfill diversion practices and carbon reduction practices, ranging from $0.40 to $1.37 per six-pack and $0.67 to $1.21 per six-pack, respectively. </a>These results indicate consumers do in fact place value on beer produced using sustainable practices, and the demographics of consumers with the greatest WTP are similar to that of craft beer consumer.</p>

The Water Framework Directive requires reduced environmental impacts from human activities and for the assessment of the non-market benefits of pollution remediation schemes. This policy shift has exacerbated the research problems surrounding the physical, social and economic consequences of the relationship between land use and water quality. This research seeks to quantify the major socio-economic and environmental benefits for people which may arise as riverine pollution is reduced. To achieve these aims this research integrates primary data analyses combining choice experiment techniques with geographical information system based analyses of secondary data concerning the spatial distributions of riverine pollution. Current knowledge on the microbial quality of river water, measured by faecal indicator organism (FIO) concentrations and assessed at catchment scale, is inadequate. This research develops generic regression models to predict base- and high-flow faecal coliform (FC) and enterococci (EN) concentrations, using land cover and population (human and livestock) variables. The resulting models are then used both to predict FIO concentrations in unmonitored watercourses and to evaluate the likely impacts of different land use scenarios, enabling insights into the optimal locations and cost-effective mix of implementation strategies. Valuation experiments frequently conflate respondents’ preferences for different aspects of water quality. This analysis uses stated preference techniques to disaggregate the values of recreation and ecological attributes of water quality, thereby allowing decision makers to better understand the consequences of adopting alternative investment strategies which favour either ecological, recreational or a mix of benefits. The results reveal heterogeneous preferences across society; specifically, latent class analysis identifies three distinct groups, holding significantly different preferences for water quality. From a methodological perspective this research greatly enhances the ongoing synthesis of geographic and economic social sciences and addresses important policy questions which are of interest to a variety of stakeholders, including government departments and the water industry.

No description available.

In this research consumers attitudes towards general food safety and their perceptions of the safety of beef in Japan and Canada are examined. Risk perceptions, the willingness to pay for beef traceability from farm to final consumer and the willingness to pay for animal testing for bovine spongifrom encephalopothy (BSE) are measured through a stated preference exercise, provided as part of national surveys in each country. Japanese respondents continue to have higher risk attitudes and perceptions about beef than Canadian respondents in 2009 as compared to 2006. In each country survey respondents strongly prefer domestic beef over imports from any other country. However, interest in beef from other countries increases as full traceability, or one hundred % animal testing for BSE or both attributes are incorporated into the markets. The willingness to pay increases at a diminishing rate, from either traceability or BSE animal testing to both attributes. In latent class models the Japanese data suggest that there are three distinct classes of survey respondents, where class 1 respondents are characterized as being more trusting and willing to pay for beef from different countries, class 2 respondents strongly prefer domestic beef and their willingness to pay for imported beef does not increase with traceability or animal testing and class 3 respondents would only be willing to pay for traceable and a combination of traceable and animal tested domestic beef. Similarly, Canadian survey respondents can be segregated into two classes. Class 1 consumers are more trusting and will be willing to pay for both domestic and imported beef. Class 2 consumers are more cautious. / Agricultural and Resource Economics

Objectif : La bronchopneumonie infectieuse (BPI) est une affection des voies respiratoires inférieures due à l’interaction entre des agents microbiens, des facteurs environnementaux et des facteurs propres à l’individu. Malgré des années de recherche, la BPI reste prévalente en élevage de génisses de remplacement et de veaux lourds engendrant des pertes économiques majeures et une forte consommation d’antibiotiques dans ces élevages. Le score clinique respiratoire diagnostique (SCRD), l’auscultation thoracique, l’échographie thoracique et la radiographie thoracique sont des tests couramment utilisés pour détecter la BPI. Cependant, leurs performances diagnostiques actuelles à savoir : leurs facultés à détecter les veaux malades (sensibilité; Se) et leurs facultés à détecter les veaux non malades (spécificité; Sp) sont sous-optimales pour détecter adéquatement les veaux de race laitière nécessitant d’être traités avec des antibiotiques et/ou des anti-inflammatoires. L’objectif général de cette thèse était donc d’améliorer les stratégies de détection du statut actif de la BPI, statut nécessitant un traitement (d’antibiotiques et/ou d’anti-inflammatoires), afin, d’une part, de mieux diagnostiquer les veaux malades, diminuer les impacts économiques de la maladie et améliorer le bien-être des veaux ; et d’autre part, favoriser l’emploi judicieux des antibiotiques en traitant uniquement les veaux qui en ont besoin. Pour répondre à cet objectif général, 3 sous-objectifs spécifiques ont été définis : (1) développer un SCRD, en identifiant et validant les signes cliniques avec une bonne répétabilité interopérateurs pour diagnostiquer le statut actif de la BPI à l’échelle individuelle et populationnelle; (2) Améliorer les stratégies diagnostiques des tests d’imagerie médicale en comparant l’échographie et la radiographie thoracique pour détecter les lésions pulmonaires et le statut actif de la BPI; (3) Améliorer l’échographie thoracique en standardisant la technique (ajout ou non du lobe crânial droit) et l’interprétation (seuil de profondeur) afin d’en optimiser ces performances pour détecter le statut actif de la BPI. Méthodes : (1) 800 veaux lourds de 80 lots ont été filmés à l’aide d’une caméra portative lors de leur évaluation clinique. Différentes séquences vidéo ont été aléatoirement sélectionnées et soumises à un panel d’évaluateur constitué de producteurs de veaux lourds, techniciens et médecins vétérinaires. La variabilité interopérateurs de chaque signe clinique a été évaluée. Les signes cliniques ayant la meilleure répétabilité ont été retenus pour le développement et la validation d’un SCRD en utilisant un modèle bayésien d’analyse de classe latente. Enfin, des stratégies diagnostiques des BPI au niveau populationnel ont été développées; (2) 50 veaux hospitalisés ont reçu une échographie et une radiographie thoracique. La tomodensitométrie thoracique (CT scan) a été utilisé comme gold standard pour vérifier les individus positifs à l’un des deux tests. Les performances des deux tests ont été comparées pour détecter d’abord les lésions pulmonaires en comparant avec les résultats du CT scan, puis le statut actif de la BPI défini par trois experts; (3) la Se et la Sp pour diagnostiquer le statut actif de BPI de différents sites d’échographie thoracique et différents seuils de profondeur de consolidation pulmonaire ont été évalués à l’aide de modèles utilisant une analyse bayésienne à classe latente dans une population de génisse de remplacement et une population de veaux lourds. Résultats : (1) La position des oreilles et la toux induite étaient les signes cliniques de BPI les plus répétables. Le SCRD développé et validé consiste à évaluer ces deux signes cliniques ainsi que la température rectale (anormale T≥ 39.5◦C) sur 10 veaux d’un lot, deux semaines après leur arrivée en parc d’engraissement. La présence de 3 veaux avec deux de ces prédicteurs implique que le lot a 94 % de chance d’avoir une prévalence du statut actif de BPI ≥ 0.10. Avec moins de 3 veaux, le lot à 95 % de chance de ne pas avoir une prévalence ≥ 0.10; (2) Pour détecter les lésions pulmonaires, les Se et Sp de l’échographie thoracique étaient de 0,81 (intervalle de crédit bayésien à 95 % (ICB95%): 0,65; 0,92) et 0,90 (ICB95 %: 0,81; 0,96), respectivement. Les Se et Sp de la radiographie thoracique étaient de 0,86 (ICB95%: 0,62; 0,99) et 0,89 (ICB95%: 0,67; 0,99), respectivement. Pour détecter le statut actif de la BPI, les Se et Sp de l’échographie thoracique étaient de 0,84 (intervalle de confiance à 95% (IC95%): 0,60; 0,97) et 0,74 (IC95 %: 0,57; 0,86), respectivement. La Se et Sp de la radiographie thoracique étaient de 0,89 (IC95%: 0,67; 0,99) et 0,58 (IC95%: 0,39; 0,75), respectivement. Aucune différence n’était présente entre les deux tests pour détecter les lésions pulmonaires ou le statut actif de BPI; (3) La détection échographique de lésions de consolidation pulmonaire des sites caudaux au cœur avec une profondeur ≥ 3 cm conclut à la présence du statut actif de BPI avec une Se de 0,89 (ICB95%: 0,55; 1,00) et une Sp de 0,95 (ICB95%: 0,92; 0,98). Conclusion : Dans cette thèse, nous avons amélioré les stratégies diagnostiques pour détecter les veaux à traiter de BPI. L’utilisation de ces résultats pour élaborer des algorithmes décisionnels devrait ultimement permettre de diminuer les pertes économiques et de raffiner l’utilisation d’antibiotiques. / Objectives: Infectious bronchopneumonia (BPI) is the infection of the lower respiratory tract implying an interaction between microbial agents, environment, and host. Despite decades of research, BPI remains omnipresent in dairy and veal calves, responsible for major economic losses and antimicrobial consumption. Clinical respiratory scoring system (CRSC), lung auscultation, thoracic ultrasonography, and thoracic radiography are the most popular tests used to detect BPI in calves. However, the sensitivity (Se) (i.e., faculty to detect sick calves) and the specificity (Sp) (i.e., faculty to detect healthy calves) are suboptimal to accurately detecting dairy and veal calves to treat. Our main objective was to improve the diagnostic strategies to detect active BPI (BPI status needed a treatment) in order to: firstly, better diagnose sick calves, reduce BPI economic losses and improve calves’ welfare; secondly, reduce antimicrobial consumption by treating only calves that need it. For this purpose, three sub-objectives have been defined: (1) develop and validate a CRSC including inter-operator reliable respiratory clinical signs to detect active BPI at calf-level and group level; (2) compare thoracic ultrasonography and thoracic radiography to detect lung lesions and active BPI; (3) standardize thoracic ultrasonography technique (sites to ultrasound) and interpretation (lung consolidation depth threshold) to promote thoracic ultrasonography Se and Sp to detect active BPI. Methods: (1) 800 veal calves from 80 batches were filmed with a portative camera during their clinical exam. Videos were randomly selected and assessed by producers, technicians, and veterinarians. The reliability of each respiratory clinical sign was assessed. The most reliable respiratory clinical signs were kept in order to develop and validate a CRSC using latent class Bayesian analysis. Batch-level diagnostic strategies were developed; (2) Fifty hospitalized calves underwent thoracic ultrasonography and thoracic radiography. Thoracic tomography (CT scan) was used as a gold standard to check positive calves on one of both tests. Se and Sp of both tests were compared to detect lung lesions and active BPI defined by three experts; (3) Se and Sp of different ultrasound sites and different lung consolidation depth thresholds were estimated to detect active BPI using latent class Bayesian analysis in both dairy calves and veal calves’ populations. Results: (1) Ear droop/head tilt and induced cough were the most reliable respiratory clinical signs. The CRSC developed and validated implies assessing those clinical signs with rectal temperature (abnormal T≥ 39.5◦C) in 10 calves from a batch at two weeks after arriving at the fattening unit. Having two abnormal characteristics of those predictors in three calves implies that a batch has 94% of having an active BPI prevalence ≥ 0.10. A batch with <3 positive calves on 10 has 95% of not having an active BPI prevalence ≥ 0.10; (2) For detecting lung lesions, the Se and Sp of thoracic ultrasonography were 0.81 (95% Bayesian credible interval (95%BCI): 0.65; 0.92) and 0.90 (95%BCI: 0.81; 0.96), respectively. The Se and Sp of thoracic radiography were 0.86 (95%BCI: 0.62; 0.99) and 0.89 (95%BCI: 0.67; 0.99), respectively. For detecting active BPI, the Se and Sp of thoracic ultrasonography were 0.84 (95% confidence interval (95%CI): 0.60; 0.97) and 0.74 (95%CI: 0.57; 0.86), respectively. The Se and Sp of thoracic radiography were 0.89 (95%CI: 0.67; 0.99) and 0.58 (95%CI: 0.39; 0.75), respectively. There was no difference between both tests to detect both lung lesions and active BPI; (3) Thoracic ultrasonography of the sites caudal of the heart with a depth lung consolidation ≥ 3 cm has a Se of 0.89 (95%BCI: 0.55; 1.00) and a Sp of 0.95 (95%BCI: 0.92; 0.98). Conclusion: We improved diagnostic strategies to detect active BPI in this thesis. Using our results to elaborate decisional algorithms would reduce economic losses and antimicrobial consumption of BPI in dairy and veal calves.