Linear logistic models with relaxed assumptions (LLRA) as introduced by Fischer (1974) are a flexible tool for the measurement of change for dichotomous or polytomous responses. As opposed to the Rasch model, assumptions on dimensionality of items, their mutual dependencies and the distribution of the latent trait in the population of subjects are relaxed. Conditional maximum likelihood estimation allows for inference about treatment, covariate or trend effect parameters without taking the subjects' latent trait values into account. In this paper we will show how LLRAs based on the LLTM, LRSM and LPCM can be used to answer various questions about the measurement of change and how they can be fitted in R using the eRm package. A number of small didactic examples is provided that can easily be used as templates for real data sets. All datafiles used in this paper are available from http://eRm.R-Forge.R-project.org/.

No description available.

This dissertation focuses on issues related to fitting an optimal variance-covariance structure in multilevel linear modeling framework with two Monte Carlo simulation studies. In the first study, the author evaluated the performance of common fit statistics such as Likelihood Ratio Test (LRT), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) and a new proposed method, standardized root mean square residual (SRMR), for selecting the correct within-subject covariance structure. Results from the simulated data suggested SRMR had the best performance in selecting the optimal covariance structure. A pharmaceutical example was also used to evaluate the performance of these fit statistics empirically. The LRT failed to decide which is a better model because LRT can only be used for nested models. SRMR, on the other hand, had congruent result as AIC and BIC and chose ARMA(1,1) as the optimal variance-covariance structure. In the second study, the author adopted a first-order autoregressive structure as the true within-subject V-C structure with variability in the intercept and slope (estimating [tau]00 and [tau]11 only) and investigated the consequence of misspecifying different levels/types of the V-C matrices simultaneously on the estimation and test of significance for the growth/fixed-effect and random-effect parameters, considering the size of the autoregressive parameter, magnitude of the fixed effect parameters, number of cases, and number of waves. The result of the simulation study showed that the commonly-used identity within-subject structure with unstructured between-subject matrix performed equally well as the true model in the evaluation of the criterion variables. On the other hand, other misspecified conditions, such as Under G & Over R conditions and Generally misspecified G & R conditions had biased standard error estimates for the fixed effect and lead to inflated Type I error rate or lowered statistical power. The two studies bridged the gap between the theory and practical application in the current literature. More research can be done to test the effectiveness of proposed SRMR in searching for the optimal V-C structure under different conditions and evaluate the impact of different types/levels of misspecification with various specifications of the within- and between- level V-C structures simultaneously.

Longitudinal data, which is also known as repeated measures, has grown increasingly within the past years because of its ability to monitor change both within and between subjects. Statisticians in many fields of study have chosen this way of collecting data because it is cost effective and it minimizes the number of subjects required to produce a meaningful outcome. This thesis will explore the world of longitudinal studies to gain a thorough understanding of why this type of collecting data has grown so rapidly. This study will also describe several methods to analyze repeated measures using data collected on the behavior of both adolescent and adult rats. The question of interest is to see if there is a change in the mean response over time and if the covariates (age, bodyweight, gender, and time) influence those changes. After much testing, our data set has a positive nonlinear change in the mean response over time within the age and gender groups. Using a model that included random effects proved to be a better method than models that did not use any random effects. Taking the log of the response variable and using day as the random effect was overall a better fit for our dataset. The transformed model also showed all covariates except for age as being significant.

京都大学 / 新制・論文博士 / 博士(社会健康医学) / 乙第13422号 / 論社医博第16号 / 新制||社医||11(附属図書館) / 京都大学大学院医学研究科社会健康医学系専攻 / (主査)教授 佐藤 俊哉, 教授 藤渕 航, 教授 黒田 知宏 / 学位規則第4条第2項該当 / Doctor of Public Health / Kyoto University / DFAM

Method comparison studies consist of a unique study design aiming to examine agreement between two methods to measure a physiological or clinical parameter evaluated using continuous variables. Such physiological parameters are used by healthcare providers along with other clinical data to inform diagnoses and treatment decisions. When novel methods are proposed to measure a continuous physiological parameter, method comparison studies are needed to examine the agreement between this new method and an existing method that is used in standard clinical care. This standard method is generally considered as the gold standard measurement for a given physiological parameter. The issue of repeated measurements poses special challenges when conducting method comparison studies. Repeated measurements occur when a given individual included in a method comparison study has multiple measurements, which are inherently correlated with one another and are not independent (e.g., multiple glucose measurements carried out for the same patient throughout the day using a blood test, compared to a portable point-of-care device). The limits of agreement (LOA) method proposed by Bland & Altman has been adapted to adjust for the correlation between repeated measurements and is widely used for the analysis of method comparison studies that include repeated measurements. However, other statistical methods have been proposed as alternatives to LOA analysis to inform the analysis of method comparison studies with repeated measurements. There is a gap in the literature to inform this type of analysis, whereby no guidelines or synthesis of statistical methods that can be used as alternatives to the LOA method with repeated measurements have been published. Therefore, this thesis aimed to systematically review the existing literature to identify existing alternate statistical methods for the analysis of method comparison studies that include repeated measurements, using a scoping review framework. The findings of this scoping review were used to inform the analysis of the PREMISE (Point-of-care hemoglobin accuracy and transfusion outcomes in non-cardiac surgery) study, a large prospective observational method comparison study that included repeated measurements. The aim of the PREMISE study was to examine the agreement between frequently used point-of-care devices to measure hemoglobin (POCT-Hgb) and laboratory-measured hemoglobin (lab-Hgb) in the operative setting. To further increase the understanding of the challenges associated with the analysis of method comparison studies that include repeated measurements, the analyses pertaining to agreement were performed in the context of this thesis. The findings of the PREMISE study fill an important gap in the literature pertaining to transfusion decision-making in the operative setting, where there is a paucity of evidence on the accuracy of POCT-Hgb devices, as well as from trials and transfusion guideline.

No description available.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O presente estudo tem como tema central o ganho de peso e o consumo alimentar no período reprodutivo. A tese está dividida em três partes: i) revisão da literatura científica sobre os fatores de risco para o ganho de peso excessivo na gestação e para a sua retenção no pós-parto; ii) análise quantitativa e qualitativa do consumo alimentar (energia, nutrientes e alimentos) do período gestacional para o pós-parto; e iii) avaliação prospectiva da associação entre a dieta hiperproteica (&#8805; 1.2 g/kg) e a variação de peso pós-gestacional. A revisão da literatura apontou como principais determinantes da variação de peso no pós-parto os seguintes fatores: ganho de peso gestacional, peso pré-gestacional, dieta, atividade física, lactação, idade, escolaridade, renda, paridade e raça. As partes ii e iii correspondem a dois artigos, sendo que a população do estudo empregada nas duas análises foi baseada em uma coorte de mulheres no pós-parto atendidas no Centro Municipal de Saúde (CMS) Marcolino Candau, localizado no Município do Rio de Janeiro. Dentre as 709 mulheres convidadas para participar da pesquisa, 479 ingressaram na coorte. As mulheres foram recrutadas através dos serviços de pré-natal e na rotina pediátrica de imunização do BCG, oferecidos pelo CMS, e imediatamente após o parto, na maternidade central de referência. A coleta de dados ocorreu entre maio de 1999 e abril de 2001, sendo 15 meses de recrutamento e 9 meses de seguimento. As mulheres foram entrevistadas aos 15 dias, 2, 6 e 9 meses do pós-parto. Foram consideradas elegíveis para as duas analises as mulheres com idade entre 18 e 45 anos de idade, entrevistadas até 30 dias do pós-parto, sem doenças crônicas pré-existentes, com gravidez de 35 ou mais semanas gestacionais e sem gestação gemelar. O estudo sobre a composição da dieta do período gestacional para o pós-parto contou com 276 mulheres que responderam aos dois questionários de freqüência de consumo alimentar (QFCA) aplicados aos 15 dias e aos 6 meses do seguimento. Os resultados mostraram que as mulheres consumiram dietas mais energéticas na gravidez e aquelas que mais restringiram o consumo de energia no pós-parto, apresentaram um incremento na densidade de proteínas da dieta. As análises das medidas repetidos do peso foram realizadas para avaliar o efeito da dieta hiperproteica (&#8805; 1.2 g/kg) sobre a variação de peso no pós-parto, para as quais haviam dados de uma amostra de 430 mulheres. Os resultados mostraram que as mulheres com dieta hiperproteica perderam mais peso do que as mulheres com dieta normo ou hipoproteica (< 1.2 g/kg) (226 g/mês versus 123 g/mês). As duas analises indicam que um pequeno incremento no consumo de proteínas promove maior perda de peso. Estudos longitudinais futuros devem incluir avaliação relativa a segurança das dietas hiperproteicas no pós-parto. / The central focus of the present study is the body weight change and food intake of women during the reproductive cycle. The study constitutes the final work in the authors Doctorate degree. It was divided in three parts: i) revision of the scientific literature regarding the main determinants in the development of maternal obesity; ii) longitudinal comparability of quantity and quality of the mothers diet (energy, nutrients and food intake) during pregnancy and postpartum; and iii) prospective evaluation of the effects of a high-protein (HP) diet (&#8805; 1.2 g/kg) versus a low-protein (LP) diet (< 1.2 g/kg) on body weight loss (BWL) during postpartum. Study population was constituted of a cohort of postpartum women, who underwent four follow-up evaluations (at 0.5, 2, 6, and 9 months) following delivery. These data were obtained from May 1999 to April 2001 (15 months recruitment and 9 mo. of follow up) at a primary healthy service Marcolino Candau Municipal Health Center in Rio de Janeiro City, Brazil. Participants were recruited: 1) during routine prenatal care at the Municipal Health Center; 2) at the time of routine Bacillus Calmette-Guérin immunization of the newborn at the same health center; and 3) at the main maternity hospital in the study area, one to three days following delivery. Among 709 women that were invited to participate in the research, 479 accepted. Eligibility criteria for enrollment in the cohort were: age between 15 to 45 years, less than 30 days following the delivery on the date of the first interview (approximately at 0.5 mo), absence of chronic diseases, no history of actual twin birth, gestational age at delivery &#8805; 35 weeks, and household address within the area of the local health center. Women aged < 18 years (n = 47) were excluded from the analyses. All study protocols met the guidelines of the Research Ethics Committee, and were approved by the Center for Collective Health Studies of the Federal University of Rio de Janeiro (NESC/UFRJ). The first part of the study, namely the revision of scientific literature, indicated the main risk factors for body weight change during postpartum: gestational weight gain, pre-pregnancy weight, diet, physical activity, breastfeeding, age, schooling, income, parity, and race. The second part of the study was aimed at accessing the dietary composition of women during pregnancy and postpartum., For that end, a semi-quantitative food frequency questionnaire (FFQ) which had been previously validated (Sichieri & Everhart, 1998) was applied during interview with the participants in their first visit to the health center (at the second or third trimester of pregnancy), and at six months following delivery. Two hundred and seventy-six women took part in both interview stages. Results showed that women in their gestational period have a higher intake of energy in their diet, and that women who restrained their energy intake the most presented a higher protein density in their diet during postpartum. Prospective analysis to evaluate the effects of a HP diet on BWL considered 430 participants at the first interview. Results showed that women with a HP diet lost more weight over time than women with a LP diet (226 g/month versus 123 g/month).

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O presente estudo tem como tema central o ganho de peso e o consumo alimentar no período reprodutivo. A tese está dividida em três partes: i) revisão da literatura científica sobre os fatores de risco para o ganho de peso excessivo na gestação e para a sua retenção no pós-parto; ii) análise quantitativa e qualitativa do consumo alimentar (energia, nutrientes e alimentos) do período gestacional para o pós-parto; e iii) avaliação prospectiva da associação entre a dieta hiperproteica (&#8805; 1.2 g/kg) e a variação de peso pós-gestacional. A revisão da literatura apontou como principais determinantes da variação de peso no pós-parto os seguintes fatores: ganho de peso gestacional, peso pré-gestacional, dieta, atividade física, lactação, idade, escolaridade, renda, paridade e raça. As partes ii e iii correspondem a dois artigos, sendo que a população do estudo empregada nas duas análises foi baseada em uma coorte de mulheres no pós-parto atendidas no Centro Municipal de Saúde (CMS) Marcolino Candau, localizado no Município do Rio de Janeiro. Dentre as 709 mulheres convidadas para participar da pesquisa, 479 ingressaram na coorte. As mulheres foram recrutadas através dos serviços de pré-natal e na rotina pediátrica de imunização do BCG, oferecidos pelo CMS, e imediatamente após o parto, na maternidade central de referência. A coleta de dados ocorreu entre maio de 1999 e abril de 2001, sendo 15 meses de recrutamento e 9 meses de seguimento. As mulheres foram entrevistadas aos 15 dias, 2, 6 e 9 meses do pós-parto. Foram consideradas elegíveis para as duas analises as mulheres com idade entre 18 e 45 anos de idade, entrevistadas até 30 dias do pós-parto, sem doenças crônicas pré-existentes, com gravidez de 35 ou mais semanas gestacionais e sem gestação gemelar. O estudo sobre a composição da dieta do período gestacional para o pós-parto contou com 276 mulheres que responderam aos dois questionários de freqüência de consumo alimentar (QFCA) aplicados aos 15 dias e aos 6 meses do seguimento. Os resultados mostraram que as mulheres consumiram dietas mais energéticas na gravidez e aquelas que mais restringiram o consumo de energia no pós-parto, apresentaram um incremento na densidade de proteínas da dieta. As análises das medidas repetidos do peso foram realizadas para avaliar o efeito da dieta hiperproteica (&#8805; 1.2 g/kg) sobre a variação de peso no pós-parto, para as quais haviam dados de uma amostra de 430 mulheres. Os resultados mostraram que as mulheres com dieta hiperproteica perderam mais peso do que as mulheres com dieta normo ou hipoproteica (< 1.2 g/kg) (226 g/mês versus 123 g/mês). As duas analises indicam que um pequeno incremento no consumo de proteínas promove maior perda de peso. Estudos longitudinais futuros devem incluir avaliação relativa a segurança das dietas hiperproteicas no pós-parto. / The central focus of the present study is the body weight change and food intake of women during the reproductive cycle. The study constitutes the final work in the authors Doctorate degree. It was divided in three parts: i) revision of the scientific literature regarding the main determinants in the development of maternal obesity; ii) longitudinal comparability of quantity and quality of the mothers diet (energy, nutrients and food intake) during pregnancy and postpartum; and iii) prospective evaluation of the effects of a high-protein (HP) diet (&#8805; 1.2 g/kg) versus a low-protein (LP) diet (< 1.2 g/kg) on body weight loss (BWL) during postpartum. Study population was constituted of a cohort of postpartum women, who underwent four follow-up evaluations (at 0.5, 2, 6, and 9 months) following delivery. These data were obtained from May 1999 to April 2001 (15 months recruitment and 9 mo. of follow up) at a primary healthy service Marcolino Candau Municipal Health Center in Rio de Janeiro City, Brazil. Participants were recruited: 1) during routine prenatal care at the Municipal Health Center; 2) at the time of routine Bacillus Calmette-Guérin immunization of the newborn at the same health center; and 3) at the main maternity hospital in the study area, one to three days following delivery. Among 709 women that were invited to participate in the research, 479 accepted. Eligibility criteria for enrollment in the cohort were: age between 15 to 45 years, less than 30 days following the delivery on the date of the first interview (approximately at 0.5 mo), absence of chronic diseases, no history of actual twin birth, gestational age at delivery &#8805; 35 weeks, and household address within the area of the local health center. Women aged < 18 years (n = 47) were excluded from the analyses. All study protocols met the guidelines of the Research Ethics Committee, and were approved by the Center for Collective Health Studies of the Federal University of Rio de Janeiro (NESC/UFRJ). The first part of the study, namely the revision of scientific literature, indicated the main risk factors for body weight change during postpartum: gestational weight gain, pre-pregnancy weight, diet, physical activity, breastfeeding, age, schooling, income, parity, and race. The second part of the study was aimed at accessing the dietary composition of women during pregnancy and postpartum., For that end, a semi-quantitative food frequency questionnaire (FFQ) which had been previously validated (Sichieri & Everhart, 1998) was applied during interview with the participants in their first visit to the health center (at the second or third trimester of pregnancy), and at six months following delivery. Two hundred and seventy-six women took part in both interview stages. Results showed that women in their gestational period have a higher intake of energy in their diet, and that women who restrained their energy intake the most presented a higher protein density in their diet during postpartum. Prospective analysis to evaluate the effects of a HP diet on BWL considered 430 participants at the first interview. Results showed that women with a HP diet lost more weight over time than women with a LP diet (226 g/month versus 123 g/month).

The classification of observations based on repeated measurements performed on the same subject over a given period of time or under different conditions is a common procedure in many disciplines such as medicine, psychology and environmental studies. In this thesis repeated measurements follow the Growth Curve model and are classified using linear discriminant analysis. The aim of this thesis is both to examine the effect of missing data on classification accuracy and to examine the effect of additional data on classification robustness. The results indicate that an increasing amount of missing data leads to a progressive decline in classification accuracy. With regard to the effect of additional data on classification robustness the results show a less predictable effect which can only be characterised as a general tendency towards improved robustness.