The generalized log-gamma mixture model with covariates: local influence and residual analysis

In a sample of censored survival times, the presence of an immune proportion of individuals who are not subject to death, failure or relapse, may be indicated by a relatively high number of individuals with large censored survival times. In this paper the generalized log-gamma model is modified for the possibility that long-term survivors may be present in the data. The model attempts to separately estimate the effects of covariates on the surviving fraction, that is, the proportion of the population for which the event never occurs. The logistic function is used for the regression model of the surviving fraction. Inference for the model parameters is considered via maximum likelihood. Some influence methods, such as the local influence and total local influence of an individual are derived, analyzed and discussed. Finally, a data set from the medical area is analyzed under the log-gamma generalized mixture model. A residual analysis is performed in order to select an appropriate model.

Prediction in Multilevel Logistic Regression

The purpose of this article is to present a new method to predict the response variable of an observation in a new cluster for a multilevel logistic regression. The central idea is based on the empirical best estimator for the random effect. Two estimation methods for multilevel model are compared: penalized quasi-likelihood and Gauss-Hermite quadrature. The performance measures for the prediction of the probability for a new cluster observation of the multilevel logistic model in comparison with the usual logistic model are examined through simulations and an application.

The full Bayesian significance test for mixture models: results in gene expression clustering

Gene clustering is a useful exploratory technique to group together genes with similar expression levels under distinct cell cycle phases or distinct conditions. It helps the biologist to identify potentially meaningful relationships between genes. In this study, we propose a clustering method based on multivariate normal mixture models, where the number of clusters is predicted via sequential hypothesis tests: at each step, the method considers a mixture model of m components (m = 2 in the first step) and tests if in fact it should be m - 1. If the hypothesis is rejected, m is increased and a new test is carried out. The method continues (increasing m) until the hypothesis is accepted. The theoretical core of the method is the full Bayesian significance test, an intuitive Bayesian approach, which needs no model complexity penalization nor positive probabilities for sharp hypotheses. Numerical experiments were based on a cDNA microarray dataset consisting of expression levels of 205 genes belonging to four functional categories, for 10 distinct strains of Saccharomyces cerevisiae. To analyze the method's sensitivity to data dimension, we performed principal components analysis on the original dataset and predicted the number of classes using 2 to 10 principal components. Compared to Mclust (model-based clustering), our method shows more consistent results.

Improvements on ICA mixture models for image pre-processing and segmentation

Today several different unsupervised classification algorithms are commonly used to cluster similar patterns in a data set based only on its statistical properties. Specially in image data applications, self-organizing methods for unsupervised classification have been successfully applied for clustering pixels or group of pixels in order to perform segmentation tasks. The first important contribution of this paper refers to the development of a self-organizing method for data classification, named Enhanced Independent Component Analysis Mixture Model (EICAMM), which was built by proposing some modifications in the Independent Component Analysis Mixture Model (ICAMM). Such improvements were proposed by considering some of the model limitations as well as by analyzing how it should be improved in order to become more efficient. Moreover, a pre-processing methodology was also proposed, which is based on combining the Sparse Code Shrinkage (SCS) for image denoising and the Sobel edge detector. In the experiments of this work, the EICAMM and other self-organizing models were applied for segmenting images in their original and pre-processed versions. A comparative analysis showed satisfactory and competitive image segmentation results obtained by the proposals presented herein. (C) 2008 Published by Elsevier B.V.

Determinantes individuais da utilização de serviços odontológicos por adultos e idosos de baixa renda

A utilização de serviços odontológicos resulta da interação de determinantes biológicos com fatores socioculturais, familiares e comunitários, bem como de características dos sistemas de saúde. O objetivo deste estudo foi identificar os fatores individuais associados à utilização de serviços odontológicos por parte de adultos e idosos de baixa renda residentes na área de abrangência da Estratégia Saúde da Família, em Ponta Grossa, PR. A amostra constou de 246 indivíduos, com 35 anos de idade ou mais, que responderam a um questionário sobre condições socioeconômicas, necessidade percebida e acesso a serviços odontológicos. A análise dos dados foi realizada por meio de regressão logística, segundo referencial teórico baseado no Modelo Comportamental de Andersen, considerando a consulta odontológica não recente como variável dependente. Verificou-se elevada prevalência de problemas bucais auto-referidos e de perdas dentárias. Cerca de 40% dos adultos e 67% dos idosos não iam ao dentista há mais de três anos. Indivíduos que não residiam em domicílios próprios, realizavam higiene bucal com menor frequência e utilizavam próteses totais apresentaram maiores chances de haver utilizado os serviços odontológicos há mais tempo. O fato de possuir um dentista regular foi identificado como fator de proteção na análise. Concluindo, os determinantes individuais mostraram-se importantes indicadores de acesso aos serviços de saúde bucal. O modelo teórico confirmou a presença de desigualdades sociais e psicossociais na utilização de serviços odontológicos entre os adultos e idosos de baixa renda.

Clinical risk factors for osteoporotic fractures in Brazilian women and men: the Brazilian Osteoporosis Study (BRAZOS)

The Brazilian Osteoporosis Study (BRAZOS) is the first epidemiological study carried out in a representative sample of Brazilian men and women aged 40 years or older. The prevalence of fragility fractures is about 15.1% in the women and 12.8% in the men. Moreover, advanced age, sedentarism, family history of hip fracture, current smoking, recurrent falls, diabetes mellitus and poor quality of life are the main clinical risk factors associated with fragility fractures. The Brazilian Osteoporosis Study (BRAZOS) is the first epidemiological study carried out in a representative sample of Brazilian men and women aged 40 years or older with the purpose of identifying the prevalence and the main clinical risk factors (CRF) associated with osteoporotic fracture in our population. A total of 2,420 individuals (women, 70%) from 150 different cities in the five geographic regions in Brazil, and all different socio-economical classes were selected to participate in the present survey. Anthropometrical data as well as life habits, fracture history, food intake, physical activity, falls and quality of life were determined by individual quantitative interviews. The representative sampling was based on Brazilian National data provided by the 2000 and 2003 census. Low trauma fracture was defined as that resulting of a fall from standing height or less in individuals 50 years or older at specific skeletal sites: forearm, femur, ribs, vertebra and humerus. Sampling error was 2.2% with 95% confidence intervals. Logistic regression analysis models were designed having the fragility fracture as the dependent variable and all other parameters as the independent variable. Significance level was set as p < 0.05. The average of age, height and weight for men and women were 58.4 +/- 12.8 and 60.1 +/- 13.7 years, 1.67 +/- 0.08 and 1.56 +/- 0.07 m and 73.3 +/- 14.7 and 64.7 +/- 13.7 kg, respectively. About 15.1% of the women and 12.8% of the men reported fragility fractures. In the women, the main CRF associated with fractures were advanced age (OR = 1.6; 95% CI 1.06-2.4), family history of hip fracture (OR = 1.7; 95% CI 1.1-2.8), early menopause (OR = 1.7; 95% CI 1.02-2.9), sedentary lifestyle (OR = 1.6; 95% CI 1.02-2.7), poor quality of life (OR = 1.9; 95% CI 1.2-2.9), higher intake of phosphorus (OR = 1.9; 95% CI 1.2-2.9), diabetes mellitus (OR = 2.8; 95% CI 1.01-8.2), use of benzodiazepine drugs (OR = 2.0; 95% CI 1.1-3.6) and recurrent falls (OR = 2.4; 95% CI 1.2-5.0). In the men, the main CRF were poor quality of life (OR = 3.2; 95% CI 1.7-6.1), current smoking (OR = 3.5; 95% CI 1.28-9.77), diabetes mellitus (OR = 4.2; 95% CI 1.27-13.7) and sedentary lifestyle (OR = 6.3; 95% CI 1.1-36.1). Our findings suggest that CRF may contribute as an important tool to identify men and women with higher risk of osteoporotic fractures and that interventions aiming at specific risk factors (quit smoking, regular physical activity, prevention of falls) may help to manage patients to reduce their risk of fracture.

Residuals for log-Burr XII regression models in survival analysis

In this paper, we compare three residuals to assess departures from the error assumptions as well as to detect outlying observations in log-Burr XII regression models with censored observations. These residuals can also be used for the log-logistic regression model, which is a special case of the log-Burr XII regression model. For different parameter settings, sample sizes and censoring percentages, various simulation studies are performed and the empirical distribution of each residual is displayed and compared with the standard normal distribution. These studies suggest that the residual analysis usually performed in normal linear regression models can be straightforwardly extended to the modified martingale-type residual in log-Burr XII regression models with censored data.

Log-Burr XII regression models with censored data

In survival analysis applications, the failure rate function may frequently present a unimodal shape. In such case, the log-normal or log-logistic distributions are used. In this paper, we shall be concerned only with parametric forms, so a location-scale regression model based on the Burr XII distribution is proposed for modeling data with a unimodal failure rate function as an alternative to the log-logistic regression model. Assuming censored data, we consider a classic analysis, a Bayesian analysis and a jackknife estimator for the parameters of the proposed model. For different parameter settings, sample sizes and censoring percentages, various simulation studies are performed and compared to the performance of the log-logistic and log-Burr XII regression models. Besides, we use sensitivity analysis to detect influential or outlying observations, and residual analysis is used to check the assumptions in the model. Finally, we analyze a real data set under log-Buff XII regression models. (C) 2008 Published by Elsevier B.V.

Association of dental enamel lead levels with risk factors for environmental exposure

OBJECTIVE: To analyze household risk factors associated with high lead levels in surface dental enamel. METHODS: A cross-sectional study was conducted with 160 Brazilian adolescents aged 14-18 years living in poor neighborhoods in the city of Bauru, southeastern Brazil, from August to December 2008. Body lead concentrations were assessed in surface dental enamel acid-etch microbiopsies. Dental enamel lead levels were measured by graphite furnace atomic absorption spectrometry and phosphorus levels were measured by inductively coupled plasma optical emission spectrometry. The parents answered a questionnaire about their children's potential early (05 years old) exposure to well-known lead sources. Logistic regression was used to identify associations between dental enamel lead levels and each environmental risk factor studied. Social and familial covariables were included in the models. RESULTS: The results suggest that the adolescents studied were exposed to lead sources during their first years of life. Risk factors associated with high dental enamel lead levels were living in or close to a contaminated area (OR = 4.49; 95% CI: 1.69;11.97); and member of the household worked in the manufacturing of paints, paint pigments, ceramics or batteries (OR = 3.43; 95% CI: 1.31;9.00). Home-based use of lead-glazed ceramics, low-quality pirated toys, anticorrosive paint on gates and/or sale of used car batteries (OR = 1.31; 95% CI: 0.56;3.03) and smoking (OR = 1.66; 95% CI: 0.52;5.28) were not found to be associated with high dental enamel lead levels. CONCLUSIONS: Surface dental enamel can be used as a marker of past environmental exposure to lead and lead concentrations detected are associated to well-known sources of lead contamination.

Saúde gengival de adolescentes e a utilização de serviços odontológicos, Estado de São Paulo

OBJETIVO: Avaliar a associação de condições de saúde gengival com a utilização de serviço odontológico. MÉTODOS: Realizou-se levantamento epidemiológico de saúde bucal de 1.799 adolescentes, em 35 cidades do Estado de São Paulo, em 2002. A saúde gengival foi avaliada pela prevalência de sangramento na gengiva à sondagem e cálculo dentário (índice periodontal comunitário) e oclusão dentária (índice de estética dentária). A utilização de serviços odontológicos foi medida pelo índice de cuidado (O/CPO) para cada cidade. Análise multinível de regressão logística ajustou modelos explicativos para fatores associados aos desfechos de interesse. RESULTADOS: A prevalência de sangramento gengival à sondagem foi 21,5%; de cálculo dentário foi 19,4%. Os participantes do sexo masculino, negros e pardos, moradores em áreas rurais, residentes em domicílios aglomerados e com atraso escolar apresentaram chance significantemente mais elevada para os agravos que seus respectivos pares de comparação. Características de oclusão dentária também associaram com gengiva não-saudável: apinhamento dos segmentos incisais, mordida aberta vertical anterior, relação molar antero-posterior. Cidades com maior utilização de serviço odontológico tiveram menor proporção de adolescentes com sangramento gengival e cálculo. CONCLUSÕES: A utilização de serviços odontológicos foi significativamente associada a melhores condições de saúde gengival (sangramento e cálculo). Essa associação independeu das características sociodemográficas individuais e contextuais, e de oclusão dentária.

Factors associated with stages of change for red meat and vegetable intake by Japanese-Brazilians

Stages of change assess individual motivation for lifestyle changes, contributing to the development of more effective intervention strategies. The objective of the present study was to identify factors associated with stages of change for lower intake of red meat and higher intake of vegetables in a cross-sectional analysis of 578 Japanese-Brazilians aged 30-90 years. In adjusted logistic regression models, the odds ratios for women (OR = 1.89; 95%CI: 1.154; 3.103) and physically active individuals (OR = 1.00; 95%CI: 1.000; 1.001) were positively associated with stage of "action" for the higher intake of vegetables. Inverse associations were observed between central obesity (OR = 0.5; 95%CI: 0.351; 0.887) and highest tertile of red meat intake (OR = 0.50; 95%CI: 0.302; 0.817), as well as a positive association between age (OR = 1.04; 95%CI: 1.020; 1.070) and the stage of "action" to the lower intake of meat were verified. Motivation for Japanese-Brazilians to change their food intake was linked to lifestyle. Stage of change is an important factor in mediating food intake behavior change.

Associação entre prática de atividade física, escolaridade e perfil alimentar de motoristas de caminhão

A demanda da sociedade moderna intensificou o trabalho ininterrupto em diversas profissões. Além disso, o modo de organização da sociedade, com atividades predominantemente mecanizadas, tem contribuído para a prevalência de hábitos de vida não saudáveis, como a inatividade física. Este estudo objetiva verificar se fatores ocupacionais, sociodemográficos, antropométricos e alimentares estão associados à prática de atividade física insuficiente e se há diferenças nessa associação entre motoristas de caminhão que trabalham de dia ou à noite. Participaram da pesquisa 470 motoristas de caminhão, que responderam a questionários de dados sociodemográficos, atividade física e frequência alimentar. Foi realizada uma regressão logística univariada para verificar a associação entre atividade física insuficiente e as demais variáveis. Além disso, a regressão logística múltipla foi testada para obter modelos que mostrem a associação de conjuntos de variáveis relacionados à atividade física insuficiente. Os resultados indicaram que a prática de atividade física está associada ao maior nível de escolaridade (OR = 1,84; IC = 1,22-2,76) e menor consumo de bebidas alcoólicas (OR = 1,59; IC = 1,04-2,45). Maior ingestão de cereais integrais (OR = 1,63; IC = 1,08-2,46) foi associada à prática regular de atividade física. Entre os trabalhadores noturnos, foi encontrada associação entre a prática regular de atividade física, maior consumo de cereais integrais (OR = 2,02; IC =1,13-3,60) e menor consumo de carboidratos simples (OR = 1,91; IC = 1,08-3,37).

Practice of walking and its association with perceived environment among elderly Brazilians living in a region of low socioeconomic level

Background: The aim of this study was analyze associations between the practice of walking and environmental perception among elderly Brazilians in a region of low socioeconomic level. Methods: A cross-sectional study was conducted among 385 elderly people aged 60 years and over. To evaluate walking, the International Physical Activity Questionnaire (IPAQ), long version (leisure and transport modules) was used. The environment was evaluated by means of the Neighborhood Environmental Walkability Scale (NEWS) (adapted Brazilian version). For the statistical analysis, multiple logistic regression models were created separately for men and women. The practice of at least 150 minutes a week of walking was the dependent variable, and the variables of environmental perception were the independent variables. All the models were controlled for schooling level and age. Results: The proportion of elderly people active in walking was 56.9% for the men and 26.4% for the women. The perception of the presence of soccer fields (OR = 4.12) and their proximity, within ten minutes' walk from home (OR = 3.43), were associated with the practice of walking among the men. The perception of the presence of public squares (OR = 4.70) and the proximity of primary healthcare units, within ten minutes' walk from home (OR = 3.71), were associated with the practice of walking among the women. An association with adequate perception of vehicle traffic remained at the threshold of significance for the women. Conclusion: Accessibility of leisure structures such as football fields and public squares and of health services such as primary healthcare units were important environmental variables associated with the practice of walking among elderly people living in a region of low socioeconomic level in Brazil. These variables need to be taken into consideration when aiming to promote the practice of walking among elderly people living in similar regions.

Subclinical hyperthyroidism and dementia: the Sao Paulo Ageing & Health Study (SPAH)

Background: Several epidemiologic studies have shown a possible association between thyroid function and cognitive decline. Our aim was to evaluate the association of subclinical hyperthyroidism and dementia in a population sample of older people Methods: A cross-sectional study - Sao Paulo Ageing & Health Study (SPAH) - in a population sample of low-income elderly people >= 65 years-old to evaluate presence of subclinical thyroid disease as a risk factor for dementia. Thyroid function was assessed using thyrotropic hormone and free-thyroxine as well as routine use of thyroid hormones or antithyroid medications. Cases of dementia were assessed using a harmonized one-phase dementia diagnostic procedure by the ""10/66 Dementia Research Group"" including Alzheimer's disease and vascular dementia. Logistic regression models were used to test a possible association between subclinical hyperthyroidism and dementia. Results and discussion: Prevalence of dementia and of subclinical hyperthyroidism were respectively of 4.4% and 3.0%. After age adjustment, we found an association of subclinical hyperthyroidism and any type of dementia and vascular dementia (Odds Ratio, 4.1, 95% Confidence Interval [95% CI] 1.3-13.1, and 5.3 95% CI, 1.1-26.4; respectively). Analyzing data by gender, we found an association of subclinical hyperthyroidism with dementia and Alzheimer's disease only for men (OR, 8.0; 95% CI, 1.5-43.4; OR, 12.4; 95% CI, 1.2-128.4; respectively). No women with subclinical hypothyroidism presented Alzheimer's disease in the sample. Conclusion: The results suggest a consistent association among people with subclinical hyperthyroidism and dementia.

Oral Health Promotion Through an Online Training Program for Medical Students

The objective of this study was to compare the impact on knowledge and counseling skills of face-to-face and Internet-based oral health training programs on medical students. Participants consisted of 148 (82 percent) of the 180 invited students attending their fifth academic year at the Faculty of Medicine, University of Sao Paulo, Brasil, in 2007. The interventions took place during a three-month training period in the clinical Center for Health Promotion, which comprised part of a clerkship in Internal Medicine. The students were divided into four groups: 1) Control Group (Control), with basic intervention; 2) Brochure Group (Br), with basic intervention plus complete brochure with oral health themes; 3) Cybertutor Group (Cy), with basic intervention plus access to an Internet-based training program about oral health themes; and 4) Cybertutor + Contact Group (Cy+C), the same as Cy plus brief proactive contact with a tutor. The impact of these interventions on student knowledge was measured with pre- and post assessments, and student skills in asking and counseling about oral health were assessed with an objective structured clinical examination (OSCE). Multivariate logistic regression models were applied to identify the odds ratios of scoring above Control's medians on the final assessment and the OSCE. In the results, Cy+C performed significantly better than Control on both the final assessment (OR 9.4; 95% CI 2.7-32.8) and the OSCE (OR 5.6; 95% CI 1.9-16.3) and outperformed all the other groups. The Cy+C group showed the most significant increase in knowledge and the best skills in asking and counseling about oral health.