Made available in DSpace on 2011-05-04T12:42:06Z (GMT). No. of bitstreams: 0 Previous issue date: 2009 / Dados nacionais sobre perfil de doenças e utilização de serviços odontológicos são raros no Brasil. O levantamento epidemiológico SB Brasil (Saúde bucal da população brasileira), realizada entre 2002 e 2003 foi a mais abrangente pesquisa sobre as condições de saúde bucal realizada até hoje, contemplando tanto aspectos relativos às doenças bucais quanto ao uso dos serviços odontológicos. O objetivo desta tese de doutorado foi identificar fatores associados às desigualdades na utilização dos serviços odontológicos no Brasil nas faixas etárias entre 15 a 19,35 a 44 e 65 a 74 anos, à partir do banco de dados do SB Brasil 2003. No entanto, para a realização deste objetivo, foi necessário incorporar ao banco de dados deste inquérito, as informações estruturais do desenho de sua amostra, uma vez que o processo de amostragem probabilística não foi concluído. Esta tese está dividida em três artigos: O primeiro que identifica e detalha os problemas encontrados no banco de dados do inquérito SB Brasil 2003, destacando a importância do cálculo dos pesos amostrais como fundamentais para a produção de dados não enviesados em uma amostra complexa e propõe formas de correção. O segundo que descreve as atividades de recuperação das informações necessárias ao cálculo dos pesos amostrais, o método utilizado para o cálculo dos referidos pesos e a técnica de calibração dos pesos. E, um terceiro artigo que investiga os fatores associados às desigualdades na utilização dos serviços odontológicos nasfaixas etárias de 15 a 19, 35 a 44 e 65 a 74 anos, por meio de análise multivariada. Os pesos amostrais e as demais variáveis estruturais do desenho da amostra da pesquisa SB Brasil 2003 foram determinados considerando as informações disponíveis sobre a seleçãodas unidades nos municípios incluídos, e uma nova versão do banco de dados foi produzida. Análises sobre o uso dos serviços odontológicos no Brasil foram então conduzidas, com modelos explicativos da sua variação apontando desigualdades entre adultos e idosos,sobretudo. Adicionalmente, diferenças encontradas nas análises realizadas com o banco de dados original e o banco corrigido ratificam a importância do uso dos pesos amostrais para a obtenção de estimativas confiáveis, provendo informação com maior qualidade dos dados nacionais sobre a saúde bucal da população brasileira. / National data about oral health problems and use of oral health services are rare in Brazil. The epidemiological survey SB Brazil (Oral Health Conditions of the brasilian population) was held between 2002 and 2003. This survey was the most comprehensive research on oral health conditions conduced until today, contemplating aspects about oral illnesses and dental services. The aim of this thesis was to identify factors associates to the inequalities in the use of the oral health services in Brazil in three age’s groups (adolescents, adults and elderly) from the data base of research SB Brazil 2003. However, for the accomplishment of this objective, it was necessary to incorporate the data base of this research, the structural information of the drawing of its sample, a time that the process of probabilistic sampling was not concluded. This thesis is divided in three articles: The first one that it identifies and it details the problems found in the data base of SB Brazil 2003 research, detaching the importance of the calculation of the weights sampling as basic for the production of data without bias in a complex sample and to consider correction forms. The second one describes the activities of recovery of the necessary information to calculate the sampling weights, the method used for the calculation of the weights and the technique of calibration of the weights. And, one third article that investigates the factors associates to the inaqualities in the use of the dental services in the age’s groups (15-19, 35-44 and 65-74 years old) by means of a multivaried analysis. The sample weights and others structural variables of the drawing of its sample were determined considering the available information on the selection of units in the municipalities included, and a new version of data base was produced. After, analysis on the use of dental services in Brazil were then conducted, with explanatory models of its variation pointing inequalities between adults and the elderly, especially. In addition, differences found in tests carried out with the original database and the bank fixed confirm the importance of using the sample weights to obtain reliable estimates, providing information with higher quality of national data on oral health of the population.

空載光達系統於建置國土測繪基本資料扮演關鍵角色，依國土測繪法，為確保測繪成果品質，應依測量計畫目的及作業精度需求辦理儀器校正。國土測繪中心已於102年度建置航遙測感應器系統校正作業中，提出矩形建物之平屋頂面做為空載光達系統校正之可行性，而其所稱之校正，是以點雲精度評估待校件空載光達系統所得最終成果品質，並不對儀器做任何參數改正，但其校正成果可能因不同人員操作而有差異，因此本研究嘗試建立一套空載光達點雲半自動化精度評估程序，此外探討以山形屋脊線執行點雲精度評估之可行性。 由於光達點雲為離散的三維資訊，不論是以山形屋脊線或矩形建物之平屋頂面作為標物執行點雲精度評估，均須先萃取屋頂面上之點，為避免萃取成果受雜訊影響，本研究引入粗差偵測理論，發展最小一乘法結合李德仁以後驗變方估計原理導出的選擇權迭代法(李德仁法)將非屋頂點視為粗差排除。研究中分別對矩形建物之平屋頂面及山形屋脊線進行模擬及真實資料實驗，其中山形屋脊線作為點雲精度評估之可行性實驗中發現不適合用於評估點雲精度，因此後續實驗僅以萃取矩形建物之平屋頂面點雲過程探討粗差比率對半自動化點雲精度評估程序之影響。模擬實驗成果顯示最小一乘法有助於提升李德仁法偵測粗差數量5%至10%；真實資料實驗，以含有牆面點雲的狀況為例，則有助提升5%的偵測粗差數量。本研究由逐步測試結果提出能夠適用於真實狀況的半自動化之點雲精度評估程序，即使由不同人員操作，仍能獲得一致的成果，顯示本研究半自動化精度評估程序之可信度。 / The airborne LiDAR system plays a crucial role in building land surveying data. Based on the Land Surveying and Mapping Act, to ensure the quality of surveying, instrument calibration is required. The approach proposed by National Land Surveying and Mapping Center (NLSC) in 2013 was confirmed the feasibility for airborne LiDAR system calibration using rectangular horizontal roof plane. The calibration mean to assess the final quality of airborne LiDAR system based on the assessment of the accuracy of the point cloud, and do not adjust the instrument. But the results may vary according to different operators. This study attempts to establish a semi-automatic procedure for the accuracy assessment of point clouds from airborne LiDAR system. In addition, the gable roof ridge lines is discussed for its feasibility for the accuracy assessment of point cloud. No matter that calibration is performed using rectangular horizontal roof plane or gable roof ridge line, point clouds located on roof planes need to be extracted at first. Therefore, Least Absolute Deviation (LAD) combined with the Iteration using Selected Weights (Deren Li method) is developed to exclude the non-roof points which regarded as gross errors and eliminate their influences. The simulated test and actual data test found that gable roof ridge lines are not suitable for accuracy assessment. As for the simulated test using horizontal roof planes, LAD combined with Deren Li method prompts the rate of gross error detection about 5% to 10% than that only by Deren Li method. In actual test, data contains wall points, LAD combined with Deren Li method can prompt about 5%. Meanwhile, a semi-automatic procedure for real operations is proposed by the step-by-step test. Even different operators employ this semi-automatic procedure, consistent results will be obtained and the reliability can achieve.

Research in the area of public procurement has gained increasing attention with increasing interest in researching different evaluation methods for incoming tenders in procurement. The study analyzed the structure for procurement with a multi-criteria model for observation of changes in ranking and studied the importance of correct scoring. A sensitivity and robustness analysis were used as a tool to study the stability and the strength of a contracting entity's priorities. The purpose of this thesis was a study to determine if the ranking of tenders became more sensitive to the structure of the model or for each tender's evaluation, where it was found that the ability to correctly score points was more crucial than being critical of weights assigned to the criteria in the procurement, respectively. Relationship between price and quality turned out to be unknown to some extent, the connection could be determined after all tenders were submitted and accumulated points were available. / Forskning inom området offentlig upphandling har fått allt mer uppmärksamhet när intresset hamnat på undersökning av olika utvärderingsmetoder för inkommande anbud vid en upphandling. I studien analyserades uppbyggnaden för upphandlingar med en multikriteriemodell för observation över förändringar hos rangordning samt studerandet av vikten för en korrekt poängsättning. En känslighets- och robusthetsanalys tillämpades som verktyg för att studera stabiliteten och styrkan hos en upphandlande entitets prioriteringar. Syftet med uppsatsen var en undersökning för en bedömning om rangordningen av upphandlingars anbud blev känsligast mot strukturen på modellen eller för varje anbuds värdering, där det visade sig att förmågan för korrekt poängtilldelning var mer avgörande än att vara kritisk till vikterna som tilldelades respektive kriterier inom upphandlingen. Relationen mellan pris och kvalitet visade sig vara okänd till viss del, sambandet kunde avgöras efter att alla anbud var inskickade och samlade poäng fanns att tillhandahålla.

La fibrillation auriculaire (FA), l’arythmie cardiaque la plus courante est un facteur de risque majeur pour le développement de l’accident vasculaire cérébral ischémique (AVC). Les anticoagulants oraux directs (AOD) ont largement remplacé la warfarine en usage clinique pour la prévention des AVC dans la FA. Cette recherche a examiné deux défis méthodologiques importants qui peuvent survenir dans les études observationnelles sur l’efficacité et l’innocuité comparatives des AOD et de la warfarine. Premièrement : un biais d’information résultant d’une classification erronée de l’exposition au traitement à la warfarine suite aux ajustements de doses fréquentes qui ne sont pas adéquatement consignés dans les données de dispensations pharmacologiques. Deuxièmement : un biais de sélection, en raison de la censure informative, généré par des mécanismes de censure différentiels, chez les patients exposés aux AOD, ou à la warfarine. À l’aide des données administratives du Québec, j’ai mené trois études de cohortes rétrospectives qui ont portées sur toutes les personnes ayant initié un anticoagulant oral de 2010 à 2016. Ces études étaient restreintes aux résidents du Québec couverts par le régime public d'assurance médicaments (environ 40% de la population au Québec), c’est-à-dire : des personnes âgées de 65 ans et plus; des bénéficiaires de l’aide sociale; des personnes qui n’ont pas accès à une assurance-maladie privée; et les personnes à leur charge. Dans la première étude, nous avons émis l'hypothèse que les données sur les réclamations en pharmacie ne reflètent pas correctement la durée de la dispensation de la warfarine. Les écarts entre les renouvellements consécutifs étaient plus grands pour la warfarine que les AOD. Dans cette étude, on a trouvé que l'écart moyen pour les usagers de la warfarine était de 9.3 jours (avec un intervalle de confiance de 95% [IC]: 8.97-9.59), l'apixaban de 3.08 jours (IC de 95%: 2.96--3.20), et de 3.15 jours pour le rivaroxaban (IC de 95%: 3.03-3.27). Les écarts entre les renouvellements consécutifs présentaient une plus grande variabilité chez les personnes qui prenaient de la warfarine comparativement à celles qui prenaient des AOD. Cette variation peut refléter les changements de posologie de la warfarine lorsque la dose quotidienne est ajustée par le professionnel de la santé en fonction des résultats du rapport normalisé international (INR). L’ajustement de la dose peut prolonger (ou raccourcir) la période couverte par le nombre de comprimés délivrés. Dans la deuxième étude, nous avons émis l'hypothèse que la définition de la durée d'exposition basée sur la variable des « jours fournis », disponible dans la base de données, et le délai de grâce fixe, entraîneront une erreur de classification différentielle de l’exposition à la warfarine par rapport aux AOD. Dans cette étude, on a utilisé deux approches pour définir la durée des dispensations : la variable des « jours fournis » disponible dans la base de données ainsi qu’une approche axée sur les données pour la définition de la durée de dispensation qui tient compte des antécédents de distribution précédents. La deuxième étude a révélé qu'en utilisant la variable des « jours fournis », la durée moyenne (et l'écart type) des durées des dispensations pour le dabigatran, le rivaroxaban, et la warfarine étaient de 19 (15), 19 (14), et de 13 (12) jours, respectivement. En utilisant l’approche fondée sur des données, les durées étaient de 20 (16), 19 (15), et de 15 (16) jours, respectivement. Ainsi, l'approche fondée sur les données s’est rapprochée de la variable des « jours fournis » pour les thérapies à dose standard telles que le dabigatran et le rivaroxaban. Une approche axée sur les données pour la définition de la durée de dispensation, qui tient compte des antécédents de distribution précédents, permet de mieux saisir la variabilité de la durée de dispensation de la warfarine par rapport à la méthode basée sur la variable des « jours fournis ». Toutefois, cela n’a pas eu d’impact sur les estimations du rapport de risque sur la sécurité comparative des AOD par rapport à la warfarine. Dans la troisième étude, nous avons émis l'hypothèse que lors de l'évaluation de l’effet d’un traitement continu avec des anticoagulants oraux (l'analyse per-protocole), la censure élimine les patients les plus malades du groupe des AOD et des patients en meilleure santé du groupe de warfarine. Cela peut baisser l'estimation de l'efficacité et de l'innocuité comparative en faveur des AOD. L’étude a démontré que les mécanismes de censure chez les initiateurs d’AOD et de warfarine étaient différents. Ainsi, certaines covariables pronostiquement significatives, telles que l’insuffisance rénale chronique et l’insuffisance cardiaque congestive, étaient associées avec une augmentation de la probabilité de censure chez les initiateurs d’AOD, et une diminution de la probabilité de censure chez les initiateurs de warfarine. Pour corriger le biais de sélection introduit par la censure, nous avons appliqué la méthode de pondération par la probabilité inverse de censure. Deux stratégies de spécification du modèle pour l’estimation des poids de censure ont été explorées : le modèle non stratifié, et le modèle stratifié en fonction de l’exposition. L’étude a démontré que lorsque les poids de censure sont générés sans tenir compte des dynamiques de censure spécifiques, les estimés ponctuels sont biaisés de 15% en faveur des AOD par rapport à l'ajustement des estimés ponctuels avec des poids de censure stratifiée selon l’exposition (rapport de risque: 1.41; IC de 95%: 1.34, 1.48 et rapport de risque: 1.26; IC de 95%: 1.20, 1.33, respectivement). Dans l’ensemble, les résultats de cette thèse ont d’importantes implications méthodologiques pour les futures études pharmacoépidémiologiques. À la lumière de ceux-ci, les résultats des études observationnelles précédentes peuvent être revus et une certaine hétérogénéité peut être expliquée. Les résultats pourraient également être extrapolés à d’autres questions cliniques. / Atrial fibrillation (AF), the most common cardiac arrhythmia is a major risk factor for the development of ischemic stroke. Direct oral anticoagulants (DOACs) replaced warfarin in clinical use for stroke prevention in AF. This research investigated two important methodological challenges that may arise in observational studies on the comparative effectiveness and safety of DOACs and warfarin. First, an information bias resulting from misclassification of exposure to dose-varying warfarin therapy when using days supplied value recorded in pharmacy claims data. Second, a selection bias due to informative censoring with differential censoring mechanisms in the DOACs- and the warfarin exposure groups. Using the Québec administrative databases, I conducted three retrospective cohort studies that included patients initiating an oral anticoagulant between 2010 and 2016. The studies were restricted to Québec residents covered by the public drug insurance plan (about 40% of Québec’s population), including those aged 65 years and older, welfare recipients, those not covered by private medical insurance, and their dependents. In the first study, we hypothesized that pharmacy claims data inadequately captured the duration of the dispensation of warfarin. Gaps between subsequent dispensations (refill gaps) and their variation are larger for warfarin than for DOACs. In this study, we found that the average refill gap for the users of warfarin was 9.3 days (95% confidence interval [CI]:8.97-9.59), apixaban 3.08 days (95%CI: 2.96--3.20), dabigatran 3.70 days (95%CI: 3.56-3.84) and rivaroxaban 3.15 days (95%CI: 3.03-3.27). The variance of refill gaps was greater among warfarin users than among DOAC users. This variation may reflect the changes in warfarin posology when the daily dose is adjusted by a physician or a pharmacist based on previously observed international normalized ratio (INR) results. The dose adjustment may lead to a prolongation of the period covered by the number of dispensed pills. In the second study, we hypothesized that the definition of duration of dispensation based on the days supplied value and a fixed grace period will lead to differential misclassification of exposure to warfarin and DOACs. This may bias the estimate of comparative safety in favor of DOACs. In this study, we used two approaches to define the duration of dispensations: the recorded days supplied value, and the longitudinal coverage approximation (data-driven) that may account for individual variation in drug usage patterns. The second study found that using the days supplied, the mean (and standard deviation) dispensation durations for dabigatran, rivaroxaban, and warfarin were 19 (15), 19 (14), and 13 (12) days, respectively. Using the data-driven approach, the durations were 20 (16), 19 (15), and 15 (16) days, respectively. Thus, the data-driven approach closely approximated the recorded days supplied value for the standard dose therapies such as dabigatran and rivaroxaban. For warfarin, the data-driven approach captured more variability in the duration of dispensations compared to the days supplied value, which may better reflect the true drug-taking behavior of warfarin. However, this did not impact the hazard ratio estimates on the comparative safety of DOACs vs. warfarin. In the third study, we hypothesized that when assessing the effect of continuous treatment with oral anticoagulants (per-protocol effect), censoring removes sicker patients from the DOACs group and healthier patients from the warfarin group. This may bias the estimate of comparative effectiveness and safety in favor of DOACs. The study showed that the mechanisms of censoring in the DOAC and the warfarin exposure groups were different. Thus, prognostically meaningful covariates, such as chronic renal failure and congestive heart failure, had an opposite direction of association with the probability of censoring in the DOACs and warfarin groups. To correct the selection bias introduced by censoring, we applied the inverse probability of censoring weights. Two strategies for the specification of the model for the estimation of censoring weights were explored: exposure-unstratified and exposure-stratified. The study found that exposure-unstratified censoring weights did not account for the differential mechanism of censoring across the treatment group and failed to eliminate the selection bias. The hazard ratio associated with continuous treatment with warfarin versus DOACs adjusted with exposure unstratified censoring weights was 15% biased in favor of DOACs compared to the hazard ratio adjusted with exposure-stratified censoring weights (hazard ratio: 1.41; 95% CI: 1.34, 1.48 and hazard ratio: 1.26; 95%CI: 1.20, 1.33, respectively). Overall, the findings of this thesis have important methodological implications for future pharmacoepidemiologic studies. Moreover, the results of the previous observational studies can be reappraised, and some heterogeneity can be explained. The findings can be extrapolated to other clinical questions.

<div> This thesis proposes an artificial neural network (NN) model to predict fuel consumption in heavy vehicles. The model uses predictors derived from vehicle speed, mass, and road grade. These variables are readily available from telematics devices that are becoming an integral part of connected vehicles. The model predictors are aggregated over a fixed distance traveled (i.e., window) instead of fixed time interval. It was found that 1km windows is most appropriate for the vocations studied in this thesis. Two vocations were studied, refuse and delivery trucks.</div><div><br></div><div> The proposed NN model was compared to two traditional models. The first is a parametric model similar to one found in the literature. The second is a linear regression model that uses the same features developed for the NN model.</div><div><br></div><div> The confidence level of the models using these three methods were calculated in order to evaluate the models variances. It was found that the NN models produce lower point-wise error. However, the stability of the models are not as high as regression models. In order to improve the variance of the NN models, an ensemble based on the average of 5-fold models was created. </div><div><br></div><div> Finally, the confidence level of each model is analyzed in order to understand how much error is expected from each model. The mean training error was used to correct the ensemble predictions for five K-Fold models. The ensemble K-fold model predictions are more reliable than the single NN and has lower confidence interval than both the parametric and regression models.</div>