Negative log-binomial model with optimal robust variance to estimate the prevalence ratio, in cross-sectional population studies.

BACKGROUND: Cross-sectional studies are useful for the estimation of prevalence of a particular event with concerns in specific populations, as in the case of diseases or other public health interests. Most of these studies have been carried out with binary binomial logistic regression model which estimates OR values that could be overestimated due to the adjustment of the model. Thus, the selection of the best multivariate model for cross-sectional studies is a priority to control the overestimation of the associations. METHODS: We compared the precision of the estimates of the prevalence ratio (PR) of the negative Log-binomial model (NLB) with Mantel-Haenszel (MH) and the regression models Cox, Log-Poisson, Log-binomial, and the OR of the binary logistic regression in population-based cross-sectional studies. The prevalence from a previous cross-sectional study carried out in Colombia about the association of mental health disorders with the consumption of psychoactive substances (e.g., cocaine, marijuana, cigarette, alcohol and risk of consumption of psychoactive substances) were used. The precision of the point estimates of the PR was evaluated for the NLB model with robust variance estimates, controlled with confounding variables, and confidence interval of 95%. RESULTS: The NLB model adjusted with robust variance showed accuracy in the measurements of crude PRs, standard errors of estimate and its corresponding confidence intervals (95%CI) as well as a high precision of the PR estimate and standard errors of estimate after the adjustment of the model by grouped age compared with the MH PR estimate. Obtained PRs and 95%CI entre NLB y MH were: cocaine consumption (2.931,IC95%: 0.723-11.889 vs. 2.913, IC95%: 0.786-12.845), marijuana consumption (3.444, IC95%: 1.856-6.391 vs. 3.407, IC95%: 1.848, 6.281), cigarette smoking (2.175,IC95%: 1.493, 3.167 vs. 2.209, IC95%: 1.518-3.214), alcohol consumption (1.243,IC95%: 1.158-1.334 vs. 1.241, IC95%: 1.157-1.332), and risk of consumption of psychoactive substances (1.086, IC95%: 1.047-1.127 vs. 1.086, IC95%: 1.047, 1.126). The NLB model adjusted with robust variance showed mayor precision when increasing the prevalence, then the other models with robust variance with respect to MH. CONCLUSIONS: The NLB model with robust variance was shown as a powerful strategy for the estimation of PRs for cross-sectional population-based studies, as high precision levels were identified for point estimators, standard errors of estimate and its corresponding confidence intervals, after the adjustment of confounding variables. In addition, it does not represent convergence issues for high prevalence cases (as it occur with the Log-binomial model) and could be considered in cases of overdispersion and with greater precision and goodness of fit than the other models with robust variance, as it was shown with the data set of the cross-sectional study used in here.

Ethnic inequalities in reproductive, maternal, newborn and child health interventions in Ecuador: A study of the 2004 and 2012 national surveys.

Background: Analysis of health inequalities by ethnicity is critical to achieving the Sustainable Development Goals. In Ecuador, similar to other Latin American countries, indigenous and afro-descendant populations have long been subject to racism, discrimination, and inequitable treatment. Although in recent years, Ecuador has made progress in health indicators, particularly those related to the coverage of Reproductive, Maternal, Neonatal and Child Health (RMNCH) interventions, little is known as to whether inequalities by ethnicity persist. Methods: Analysis was based on two nationally representative health surveys (2004 and 2012). Ethnicity was self-reported and classified into three categories (Indigenous/Afro-Ecuadorian/Mixed ancestry). Coverage data for six RMNCH health interventions were stratified for each ethnic group by level of education, area of residence and wealth quintiles. Absolute inequality measures were computed and multivariate analysis using Poisson regression was undertaken. Findings: In 2012, 74.4% of women self-identifying as indigenous did not achieve the secondary level of education and 50.7% were in the poorest quintile (Q1); this profile was relatively unchanged since 2004. From 2004 to 2012, the coverage of RMNCH interventions increased for all ethnic groups, and absolute inequality decreased. However, in 2012, regardless of education level, area of residence and wealth quintiles, ethnic inequalities remained for almost all RMNCH interventions. Indigenous women had 24% lower prevalence of modern contraceptive use (Prevalence ratio [PR] = 0.76; 95% IC: 0.7-0.8); 28% lower prevalence of antenatal care (PR = 0.72; 95% IC: 0.6-0.8); and 35% lower prevalence of skilled birth attendance and institutional delivery (PR = 0.65; 95% IC: 0.6-0.7 and PR = 0.65; 95% IC: 0.6-0.7 respectively), compared with the majority ethnic group in the country. Interpretation: While the gaps have narrowed, indigenous people in Ecuador continue in a situation of structural racism and are left behind in terms of access to RMNCH interventions. Strategies to reduce ethnic inequalities in the coverage services need to be collaboratively redesigned/co-designed. Funding: This paper was made possible with funds from the Bill & Melinda Gates Foundation [Grant Number: INV-007,594/OPP1148933].

Factors associated with the presence of the Epstein-Barr virus in the oral cavity of high school students from the city of Cali (Colombia).

INTRODUCTION: The Epstein-Barr virus (EBV) is classified as a group 1 carcinogen. The main route of EBV transmission is oral, through saliva. The present study aimed to determine the frequency of EBV detection in the oral cavity in high school students in the city of Cali (Colombia). MATERIALS AND METHODS: Analytical cross-sectional study in order to determine the frequency of EBV detection in the oral cavity, the reasons for its prevalence and its association with several factors, in 1565 individuals. The variables analyzed were sociodemographic factors, oral hygiene, oral health, sexual behavior, cigarrete smoking and alcohol intake. The association between the EBV detection and the variables evaluated was done through a generalized linear regression model with logarithmic linkage and Poisson distribution with robust variance. RESULTS: The percentage of exposure to EBV in the oral cavity was 38.40% (CI 95%: 36.02-40.84). The frequency of presenting EBV exposure was 22% higher in men and the risk increased according to sexual behaviour. An inverse association with the school grade was found: the eleventh-grade participants had 27% less frequency of exposure to EBV than the lower grades (sixth to eighth). When analyzing the logistic model to study the association between EBV detection and independent variables, the association was overestimated. The overestimation ranged from 27% to 47% depending on the type of variable. CONCLUSIONS: The frequency of EBV detection in the oral cavity of healthy students was similar to that previously described. Factors associated to sexual behavior increased the risk of opportunity to be exposed to EBV.

The burden of excessive saturated fatty acid intake attributed to ultra-processed food consumption: a study conducted with nationally representative cross-sectional studies from eight countries.

Cross-sectional nutritional survey data collected in eight countries were used to estimate saturated fatty acid intakes. Our objective was to estimate the proportion of excessive saturated fatty acid intakes (>10 % of total energy intake) that could be avoided if ultra-processed food consumption was reduced to levels observed in the first quintile of each country. Secondary analysis was performed of 24 h dietary recall or food diary/record data collected by the most recently available nationally representative cross-sectional surveys carried out in Brazil (2008-9), Chile (2010), Colombia (2005), Mexico (2012), Australia (2011-12), the UK (2008-16), Canada (2015) and the US (2015-16). Population attributable fractions estimated the impact of reducing ultra-processed food consumption on excessive saturated fatty acid intakes (above 10 % of total energy intake) in each country. Significant relative reductions in the percentage of excessive saturated fatty acid intakes would be observed in all countries if ultra-processed food consumption was reduced to levels observed in the first quintile's consumption. The reductions in excessive intakes ranged from 10⋅0 % (95 % CI 6⋅2-13⋅6 %) in Canada to 35⋅0 % (95 % CI 28⋅7-48⋅0 %) in Mexico. In all eight studied countries, all presenting more than 30 % of intakes with excessive saturated fatty acids, lowering the dietary contribution of ultra-processed foods to attainable, context-specific levels was shown to be a potentially effective way to reduce the percentage of intakes with excessive saturated fatty acids, which may play an important role in the prevention of non-communicable diseases, particularly cardiovascular diseases.

Factors associated with the presence of the Epstein-Barr virus in the oral cavity of high school students from the city of Cali (Colombia). / Factores asociados a la presencia del virus de Epstein-Barr en la cavidad oral de adolescentes de la ciudad de Cali (Colombia).

INTRODUCTION: The Epstein-Barr virus (EBV) is classified as a group 1 carcinogen. The main route of EBV transmission is oral, through saliva. The present study aimed to determine the frequency of EBV detection in the oral cavity in high school students in the city of Cali (Colombia). MATERIALS AND METHODS: Analytical cross-sectional study in order to determine the frequency of EBV detection in the oral cavity, the reasons for its prevalence and its association with several factors, in 1,565 individuals. The variables analyzed were sociodemographic factors, oral hygiene, oral health, sexual behavior, cigarrete smoking and alcohol intake. The association between the EBV detection and the variables evaluated was done through a generalized linear regression model with logarithmic linkage and Poisson distribution with robust variance. RESULTS: The percentage of exposure to EBV in the oral cavity was 38.40% (CI 95%: 36.02-40.84). The frequency of presenting EBV exposure was 22% higher in men and the risk increased according to sexual behaviour. An inverse association with the school grade was found: the eleventh-grade participants had 27% less frequency of exposure to EBV than the lower grades (sixth to eighth). When analyzing the logistic model to study the association between EBV detection and independent variables, the association was overestimated. The overestimation ranged from 27 to 47% depending on the type of variable. CONCLUSIONS: The frequency of EBV detection in the oral cavity of healthy students was similar to that previously described. Factors associated to sexual behavior increased the risk of opportunity to be exposed to EBV.

Ocorrência de distúrbio musculoesquelético em pescadoras artesanais/ marisqueiras na Baía de Todos os Santos: uma análise sobre horas dedicadas ao trabalho / Occurrence of musculoskeletal disorders in artisanal /shellfish fisherwomen in Baía de Todos os Santos: an analysis of hours dedicated to work

Introdução: os distúrbios musculoesqueléticos são caracterizados por comprometimentos inflamatórios e/ou degenerativos, em diversos segmentos corporais, e atingem diversos grupos de trabalhadores, no entanto, ainda existem poucas evidências científicas que verificam possíveis relações desses distúrbios e horas de trabalho diárias dos pescadores artesanais. Objetivo: analisar as horas diárias dedicadas ao trabalho de pescadora e a prevalência de distúrbio musculoesquelético em regiões de membros superiores (ombro) e pescoço. Metodologia: realizou-se um estudo de corte transversal, envolvendo amostra probabilística da população de pescadoras artesanais/marisqueiras do município de Saubara, totalizando 209 pessoas. Utilizaram-se instrumentos validados para população brasileira, tais como: o Nordic Musculoskeletal Questionnaire, Job Content Questionnaire e Demandas físicas adaptadas para o trabalho. Resultados: evidenciou-se a prevalência de distúrbios musculoesqueléticos em região de pescoço ou ombro, em 71,2% (n=146) das marisqueiras. Ressalta-se que não houve interação entre as variáveis horas trabalhadas/dia e idade, com o p-valor = 0,1807. Verificaram-se associações positivas entre as marisqueiras que trabalhavam mais que 11h por dia com razão de prevalência de 1,23 e intervalo de confiança 95%: [1,04; 1,45] em relação às que trabalhavam menos que 11h, bem como a razão de prevalência de 1,38 e intervalo de confiança 95%: [1,15; 1,65] nas pescadoras com idade superior a 38 em comparação com as mais jovens. Desta forma denotando a degradação da saúde com a carga exaustiva de trabalho.

Introduction: musculoskeletal disorders are characterized by inflammatory and/or degenerative impairments in several body segments and reach different groups of workers; however, there is still little scientific evidence to verify the possible relationships between MSD and daily working hours of artisanal fishermen. Objective: to analyze the daily hours dedicated to work as a fisherwoman and prevalence of musculoskeletal disorders in regions of upper limbs (shoulder) and neck. Methodology: a cross-sectional study was carried out involving a probabilistic sample of the population of artisanal/shellfish fisherwomen in the municipality of Saubara, for a total of 209 individuals. Validated instruments were used for the Brazilian population, such as: the Nordic Musculoskeletal Questionnaire, the Job Content Questionnaire and the Physical Demands adapted to this work. Results: the prevalence of musculoskeletal disorders in the neck or shoulder region was found in 71.2% (n = 146) of shell fisherwomen. It should come to attention that there was no interaction between the variables worked hours/day and age, with p-value = 0.1807. There were positive associations among the shellfish workers that worked more than 11h with the prevalence ratio of 1.23 and 95% confidence interval: [1.04; 1.45] than those who worked less than 11 hours, as well as the prevalence ratio of 1.38 and 95% confidence interval: [1.15; 1.65] in fisherwoman aged over 38 compared to younger ones. Thus denoting health degradation with an exhaustive workload.

Work Status, Financial Stress, Family Problems, and Gender Differences in the Prevalence of Depression in Chile.

INTRODUCTION: It is widely known that women have a higher prevalence of depression than men. These differences may be explained by social differences between women and men due to gender roles. In Chile, as elsewhere, women have greater household responsibilities, lower job incomes, and especially low labor market participation. However, the incidence of these gender differences on the higher prevalence of depression in women requires further study in the Chilean context. OBJECTIVE: To identify main social differences between women and men associated with gender differences in the prevalence of depression in middle-aged Chilean men and women. DATA SOURCE: Data comes from the second Chilean National Health Survey (2009-2010), a cross-sectional, nationally representative sample (n = 2771) composed of 1103 men and 1668 women (39.8 and 60.2%, respectively), whose age range goes from 25 to 55 years old. This study was approved by the Ethics Committee of Pontificia Universidad Católica de Chile. METHODS: Prevalence ratios were calculated through Poisson regression models to estimate associations between the prevalence of past episodes of depression and social stressful life events variables for men and women separately. Gender prevalence ratios of depression (Gender PR) adjusted for age and subsequently adjusted by the social and stressful life events variables. The analyses considered factors such as age, educational level, per-capita household income, work status, role as the head of household, marital status, events of violence, family problems, personal health problems or accidents, and self-reported financial stress. RESULTS: This study finds that doing housework, reporting a serious family problem and having high financial stress were associated with a higher prevalence of depression in both genders. Whereas, health problems were only associated with prevalence in men. The age-adjusted gender PR was 2.84 [confidence interval (CI): 2.0-4.1], and when all selected variables were included attenuated to a PR of 1.86 (CI 1.3-2.7). The variable most strongly associated with depression in the fully-adjusted model was housework (PR: 5.3; CI: 1.3-21.0). CONCLUSION: In conclusion, this study finds that depression in in Chile is associated with social factors such as participation in housework, family problems, and financial stress, all of which are more common in women. To make further progress in the study of this public health problem in Chile, it is essential to incorporate more detailed characterization of gender roles in surveys and other studies. Likewise, social policies and interventions that contribute to reduce gender social inequalities in the exposure to adverse life circumstances can contribute to reduce unnecessary and avoidable gender disparities in health.

Spatial-temporal and phylogeographic characterization of Trypanosoma spp. in cattle (Bos taurus) and buffaloes (Bubalus bubalis) reveals transmission dynamics of these parasites in Colombia.

Animal Trypanosomiasis (AT) is one of the most important problems in the Colombian livestock industry reducing its production around 30%. Caribbean and Orinoquia regions play a significant role in the development of this industry, having about 6.9 million cattle and 113,000 buffaloes. Considering the paucity in studies to understand the epidemiological features and control of AT in Colombia, the present study reports the seasonal transmission patterns and phylogeographic traits of the causal agents of AT in cattle and buffaloes from these regions. Between 2014 and 2016, a three-point longitudinal survey was designed to evaluate the mentioned characteristics. Molecular analysis in cattle showed an AT prevalence of 39.2% (T. theileri 38.6%, T. evansi 6.7% and T. vivax 0.2%), with higher values during wet and late wet seasons, while in buffaloes the prevalence was 28.2% (T. theileri 28.2% and T. evansi 1.3%), with higher values during the dry season. Additionally, variables such as tabanid abundance, vector control, breeding system, age and anemia signs were significantly associated with AT prevalence (P<0.05). Only T. theileri infection was higher in cattle with anemia signs than those with normal packed cell volume. Finally, phylogeographic analysis revealed that Colombian T. theileri isolates were associated to specific host genotypes IA and IIB, described worldwide; T. vivax isolates were related to the genotype from West Africa; while T. evansi isolates are related to the South American genotypes and to new genotypes. This is the first longitudinal survey that evaluates through molecular methods, the infection of Trypanosoma spp. in two important livestock regions from Colombia, showing that the clinical effects and prevalence of these trypanosomes in cattle and buffaloes are modulated by seasonal variations, host factors, and parasite traits. The results suggest that these factors have to be taken into account to successfully control AT in these regions.

Odds Ratio or Prevalence Ratio? An Overview of Reported Statistical Methods and Appropriateness of Interpretations in Cross-sectional Studies with Dichotomous Outcomes in Veterinary Medicine.

One of the most commonly observational study designs employed in veterinary is the cross-sectional study with binary outcomes. To measure an association with exposure, the use of prevalence ratios (PR) or odds ratios (OR) are possible. In human epidemiology, much has been discussed about the use of the OR exclusively for case-control studies and some authors reported that there is no good justification for fitting logistic regression when the prevalence of the disease is high, in which OR overestimate the PR. Nonetheless, interpretation of OR is difficult since confusing between risk and odds can lead to incorrect quantitative interpretation of data such as "the risk is X times greater," commonly reported in studies that use OR. The aims of this study were (1) to review articles with cross-sectional designs to assess the statistical method used and the appropriateness of the interpretation of the estimated measure of association and (2) to illustrate the use of alternative statistical methods that estimate PR directly. An overview of statistical methods and its interpretation using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines was conducted and included a diverse set of peer-reviewed journals among the veterinary science field using PubMed as the search engine. From each article, the statistical method used and the appropriateness of the interpretation of the estimated measure of association were registered. Additionally, four alternative models for logistic regression that estimate directly PR were tested using our own dataset from a cross-sectional study on bovine viral diarrhea virus. The initial search strategy found 62 articles, in which 6 articles were excluded and therefore 56 studies were used for the overall analysis. The review showed that independent of the level of prevalence reported, 96% of articles employed logistic regression, thus estimating the OR. Results of the multivariate models indicated that logistic regression was the method that most overestimated the PR. The findings of this study indicate that although there are methods that directly estimate PR, many studies in veterinary science do not use these methods and misinterpret the OR estimated by the logistic regression.

Prevalência e fatores associados à automedicação em adultos no Distrito Federal: estudo transversal de base populacional / Prevalencia y factores asociados a la automedicación en adultos en el Distrito Federal, Brasil: estudio transversal de base poblacional / Prevalence and associated factors of self-medication in adults living in the Federal District, Brazil: a cross-sectional, population-based study

OBJETIVO: estimar a prevalência e investigar fatores associados à automedicação em adultos no Distrito Federal, Brasil. MÉTODOS: estudo transversal de base populacional, com adultos selecionados por amostragem probabilística; a prevalência da automedicação foi obtida entre os que referiram ter utilizado medicamento nos últimos sete dias; foi empregado modelo de regressão de Poisson com variância robusta para estimar razões de prevalência. RESULTADOS: entrevistaram-se 1.820 pessoas, das quais 646 usaram pelo menos um medicamento; a prevalência da automedicação foi de 14,9% (IC95%: 12,6%;17,5%); a análise ajustada apontou associação negativa em pessoas na idade de 50 a 65 anos (RP=0,26; IC95%: 0,15;0,47) e com doenças crônicas (RP=0,38; IC95%: 0,28;0,51); adultos com dificuldades na prática de atividades cotidianas (RP=2,25; IC95%: 1,43;3,53) realizaram mais automedicação. CONCLUSÃO: a automedicação foi maior em adultos jovens e naqueles com dificuldades na realização de atividades cotidianas.

OBJETIVO: estimar la prevalencia de la automedicación e investigar los factores asociados en adultos del Distrito Federal, Brasil. MÉTODOS: estudio transversal de base poblacional en adultos seleccionados mediante muestreo probabilístico; para obtener la prevalencia de automedicación se consideró el uso de medicamentos en los últimos siete días; las razones de prevalencia se definieron por medio de un modelo de regresión de Poisson ajustado con varianza robusta. RESULTADOS: se entrevistaron 1.820 personas, de las cuales 646 han usado al menos un medicamento; la prevalencia de automedicación fue de 14,9% (IC95%: 12,6%;17,5%); el análisis ajustado mostró una asociación negativa en personas de 50 a 65 años (RP=0,26 - IC95%: 0,15;0,47) y enfermedades crónicas (RP=0,38 - IC95%: 0,28;0,51); los adultos con problemas en la práctica de actividades cotidianas (RP=2,25 - IC95%: 1,43;3,53) se auto medicaron más. CONCLUSIÓN: automedicación se asoció con los adultos jóvenes y los que tienen problemas para realizar las actividades diarias.

OBJECTIVE: to assess the prevalence of self-medication and to investigate its associated factors in adults living in the Federal District, Brazil. METHODS: this is a cross-sectional population-based study conducted with adults selected through probabilistic sampling; self-medication prevalence was obtained from those who reported having used any medicine in the previous seven days; Poisson regression model with robust variance was applied to adjust the prevalence ratios. RESULTS: 1,820 individuals were interviewed, of which 646 had taken at least one medicine; self-medication prevalence was of 14.9% (95%CI: 12.6%;17.5%); the adjusted analysis showed negative association in people aged 50 to 65 years (PR=0.26; 95%CI: 0.15;0.47) and with chronic diseases (PR=0.38; 95%CI: 0.28;0.51); adults with difficulties in performing daily activities (PR=2.25; 95%CI: 1.43;3.53) practiced more self-medication. CONCLUSION: self-medication was associated to young adults and those with problems in performing daily activities.

Regresiones aplicadas al estudio de eventos discretos en epidemiología / Regressions applied to the study of discrete events in epidemiology

En este manuscrito se revisan algunos aspectos básicos de la utilización de regresiones en los estudios epidemiológicos, haciendo énfasis en aquellas aplicadas al estudio de eventos discretos. De esta manera se hace una introducción a los modelos lineales generalizados, cuya estructura es una extensión de una ecuación lineal para analizar desenlaces discretos. De este modo podemos estimar medidas de asociación como la razón de tasas usando la regresión de Poisson, o bien, el riesgo relativo (o la razón de prevalencias) usando la regresión log-binomial. En cada caso es esencial conocer la naturaleza de la variable dependiente, su distribución y reconocer las limitaciones de cada una de las herramientas de análisis.

Some basic aspects about using regressions in epidemiological studies are reviewed. Particularly, this manuscript focused on those applied to the study of discrete events. Generalized lineal models, such as Poisson and log-binomial, have a structure that is an extension of a lineal equation to analyze discrete outcomes. Thus, we can estimate association measures as the incidence rate ratio, using the Poisson regression, or the relative risk (or prevalence ratio), using log-binomial regression. In each case it is essential to know the nature of the dependent variable, as well as, its distribution and recognize the limitations of each analysis tool.

Obtaining adjusted prevalence ratios from logistic regression models in cross-sectional studies / Obtendo razões de chance prevalentes de modelos de regressão logística em estudos transversais / La obtención de las prevalencias ajustadas a partir de los modelos de regresión logística en los estudios transversales

In the last decades, the use of the epidemiological prevalence ratio (PR) instead of the odds ratio has been debated as a measure of association in cross-sectional studies. This article addresses the main difficulties in the use of statistical models for the calculation of PR: convergence problems, availability of tools and inappropriate assumptions. We implement the direct approach to estimate the PR from binary regression models based on two methods proposed by Wilcosky & Chambless and compare with different methods. We used three examples and compared the crude and adjusted estimate of PR, with the estimates obtained by use of log-binomial, Poisson regression and the prevalence odds ratio (POR). PRs obtained from the direct approach resulted in values close enough to those obtained by log-binomial and Poisson, while the POR overestimated the PR. The model implemented here showed the following advantages: no numerical instability; assumes adequate probability distribution and, is available through the R statistical package.

Nas últimas décadas, tem sido discutido o uso da razão de prevalência (RP) ao invés da razão de chance como a medida de associação a ser estimada em estudos transversais. Discute-se as principais dificuldades no uso de modelos estatísticos para o cálculo da RP: problemas de convergência, disponibilidade de ferramentas e pressupostos não apropriados. O objetivo deste estudo é implementar uma abordagem direta para estimar a RP com base em modelos logísticos binários baseados em dois métodos propostos por Wilcosky & Chamblers, e comparar com outros métodos. Utilizou-se três exemplos e comparou-se as estimativas bruta e ajustada da RP obtidas pela função com as estimativas obtidas pelos modelos log-binomial, Poisson e razão de chance prevalente (RCP). As RP da abordagem proposta resultaram em valores próximos aos obtidos pelos modelos log-binomial e Poisson, e a RCP superestimou a RP. O modelo aqui implementado apresentou as seguintes vantagens: não apresenta instabilidade numérica; assume a distribuição de probabilidades adequada; e está disponível no programa estatístico R.

En las últimas décadas, se ha discutido el uso de la razón de prevalencia (RP), en lugar del odds ratio como medida de asociación que se estima en estudios transversales. Se analizan las principales dificultades en el uso de modelos estadísticos para el cálculo de la RP: problemas de convergencia, disponibilidad de herramientas y supuestos no apropiados. El objetivo es realizar un enfoque directo para estimar la RP desde modelos logísticos binarios, basados en dos métodos propuestos por Wilcosky y Chamblers y compararlos con otros métodos. Se han utilizado 3 ejemplos y comparamos las estimaciones crudas y ajustadas de RP con las estimaciones obtenidas por log-binomial, Poisson y odds ratio de prevalencia (ORP). Los RP obtenidos del enfoque directo dieron como resultado valores cercanos a los obtenidos mediante el log- binomial y de Poisson, mientras que la RCP sobreestimó la RP. El modelo que aquí se presenta implementó las siguientes ventajas: no presenta inestabilidad numérica, toma una distribución de probabilidad apropiada y está disponible en software estadístico libre R.

Prevalence and sociodemographic characteristics of women with induced abortion in a population sample of São Paulo, Brazil / Prevalência e características sociodemográficas de mulheres com aborto provocado em uma amostra da população da cidade de São Paulo, Brasil

Objectives: This study aims at estimating the prevalence of women with induced abortion among women of childbearing age (15-49 years) who had any previous pregnancy, in the city of Sao Paulo, Brazil, in the last quarter of 2008, and identifying the sociodemographic characteristics (SC) associated with it. Methods: A cross-sectional survey was carried out. The dependent variable was dichotomized as: no abortion and induced abortion. The independent variables were: age, paid work/activity, familial monthly income, schooling, marital status, contraceptive use and number of live births. Statistical analysis was performed using log-binomial regression models with approximation of Poisson to estimate the prevalance ratios (PR). Results: Of all women with any previous pregnancy (n = 683), 4.5% (n = 31) reported induced abortion. The final multivariate model showed that having now between 40 and 44 years (PR = 2.76, p = 0.0043), being single (PR = 2.79, p = 0.0159), having 5 or more live births (PR = 3.97, p = 0.0013), current oral contraception or IUD use (PR = 2.70, p = 0.454) and using a "non effective" (or of low efficacy) contraceptive method (PR = 4.18, p = 0.0009) were sociodemographic characteristics associated with induced abortion in this population. Conclusions: Induced abortion seems to be used to limit fertility, more precisely after having reached the desired number of children. The inadequate use or non-use of effective contraceptive methods, and / or the use of contraceptives " non effective", exposed also the women to the risk of unintended pregnancies and, therefore, induced abortions. In addition, when faced with a pregnancy, single women were more likely to have an abortion than married women. .

Objetivos: O presente estudo busca estimar a prevalência de mulheres com aborto provocado dentre as mulheres em idade fértil (15 a 49 anos de idade) que apresentaram alguma gestação prévia, residentes na cidade de São Paulo, Brasil, no último trimestre de 2008, bem como identificar essa ocorrência em função de algumas características sociodemográficas (CSD). Métodos: Trata-se de um estudo transversal. A variável dependente aborto foi dicotomizada em: aborto provocado (AP) e nenhum aborto (NA). As variáveis independentes utilizadas foram: idade, atividade remunerada, renda mensal familiar, escolaridade, estado conjugal, uso de método(s) contraceptivo(s) e número de filhos nascidos vivos. A análise estatística foi realizada por meio de modelos de regressão log-binomial com aproximação de Poisson para avaliar as razões de prevalência (RP). Resultados: Dentre o total de mulheres com alguma gestação prévia (n = 683), 4,5% (n = 31) declararam algum AP. O modelo log-binomial final mostrou que idade atual entre 40 e 44 anos de idade (RP = 2,76; p = 0,0043), estado conjugal solteira (RP = 2,79; p = 0,0159), número de filhos nascidos vivos igual ou maior que 5 (RP = 3,97; p = 0,0013, uso de contraceptivo "não eficaz" ou de baixa eficácia (RP = 4,18; p = 0,0009) e uso atual de contraceptivo oral ou dispositivo intrauterino (RP = 2,70; p = 0,454) foram as características que se apresentaram mais fortemente associadas à ocorrência de AP entre as mulheres dessa população. Conclusões: O aborto provocado é utilizado para limitar a fecundidade, mais precisamente após atingir-se o número desejado de filhos. O uso inadequado, ou o não-uso, de contraceptivos eficazes e/ou a utilização de contraceptivos não eficazes, ou de baixa eficácia, ...

Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives / Vies da razao de chances estimada pela regressao logistica multinomial para estimar o risco relativo ou a razao de prevalencia e alternativas / El sesgo del odds ratio estimado por la regresión logística multinomial para estimar el riesgo relativo o la razón de prevalencia y alternativas

Recent studies have emphasized that there is no justification for using the odds ratio (OR) as an approximation of the relative risk (RR) or prevalence ratio (PR). Erroneous interpretations of the OR as RR or PR must be avoided, as several studies have shown that the OR is not a good approximation for these measures when the outcome is common (> 10%). For multinomial outcomes it is usual to use the multinomial logistic regression. In this context, there are no studies showing the impact of the approximation of the OR in the estimates of RR or PR. This study aimed to present and discuss alternative methods to multinomial logistic regression based upon robust Poisson regression and the log-binomial model. The approaches were compared by simulating various possible scenarios. The results showed that the proposed models have more precise and accurate estimates for the RR or PR than the multinomial logistic regression, as in the case of the binary outcome. Thus also for multinomial outcomes the OR must not be used as an approximation of the RR or PR, since this may lead to incorrect conclusions.

Recentes trabalhos têm enfatizado que já não há justificativa para o uso da razão de chances (RC) como aproximação do risco relativo (RR) ou razão de prevalência (RP). Deve-se evitar a interpretação equivocada da RC como RR ou RP, pois vários estudos demonstraram que a RC não é uma boa aproximação para tais medidas quando o desfecho é comum (> 10%). Para desfechos multinomiais é usual aplicar a regressão logística multinomial. Nesse contexto, não há estudos demonstrando o impacto da aproximação da RC nas estimativas de RR ou RP. O objetivo deste trabalho é apresentar e discutir métodos alternativos à regressão logística multinomial, baseados na regressão de Poisson e no modelo log-binomial. As abordagens foram comparadas por um estudo de simulação com diversos cenários. Assim como no caso do desfecho binário, os modelos propostos apresentaram estimativas mais precisas e acuradas para o RR ou RP do que a regressão logística multinomial. Então, também para os desfechos multinomiais não se deve utilizar a RC como aproximação do RR ou RP, pois conclusões incorretas podem ocorrer.

Los trabajos recientes han enfatizado que ya no se justifica el uso del odds ratio (OR) como una aproximación del riesgo relativo (RR) o razón de prevalencias (RP). El OR no puede ser interpretado como RR o RP, pues varios estudios han demostrado que el OR no es una buena aproximación cuando el suceso es común (> 10%). Para sucesos multinomiales se utiliza comúnmente la regresión logística multinomial. En este contexto, no hay estudios que demuestren el impacto de la aproximación del OR en las estimaciones de RR o RP. Nuestro objetivo es presentar y discutir métodos alternativos a la regresión logística multinomial, en base a la regresión de Poisson y al modelo log-binomial. Los enfoques utilizados fueron comparados en un estudio de simulación con diferentes escenarios. Así como en el caso de suceso binario, los modelos propuestos dieron como resultado estimaciones para RR o RP más precisas y esmeradas que la regresión logística multinomial. Para sucesos multinomiales el OR tampoco debe ser utilizado como aproximación del RR o de la RP, pues se puede llegar a conclusiones incorrectas.

Factores asociados con no control metabólico en diabéticos pertenecientes a un programa de riesgo cardiovascular / Factors associated with no metabolic control in diabetics belonging to a program of cardiovascular risk

Antecedentes: se desconoce la prevalencia de la falla en el control metabólico en pacientes diabéticos, así como los factores de riesgo asociados a ello en la ciudad de Bucaramanga. Material y métodos: se realizó un estudio analítico de corte transversal, que midió la prevalencia del no control metabólico en los pacientes diabéticos tipo 2, del programa de promoción y prevención de riesgo cardiovascular en una empresa promotora de salud de Bucaramanga; el no control metabólico se midió a través de la hemoglobina glicosilada (mayor de 7.0%). se estimaron los efectos de las variables evaluadas (personales, clínicas y de percepción de la atención) por medio de razones de prevalencia (RP) y un modelo multivariable. Resultados: se incluyeron 484 pacientes, de los cuales 407 tuvieron registro de Hba1c durante los tres meses previos al estudio. la prevalencia del no control metabólico fue de 65%. las variables asociadas con este desenlace con p < de 0.05, fueron: antecedente de dislipidemia: RP 1,14 (IC 1.00-1.32), no adherencia a los fármacos: RP 1.6 (IC 1.01-1.34), tener entre 50 y 65 años: RP 1,16 (IC 1.01-1.34), consumir más de seis tabletas de medicamentos al día: RP 1.22 (IC 1.06-1.41), estar desempleado: RP 1.38 (IC 1.09-1.76), asistir acompañado a veces a la consulta: RP 1.29 (IC 1.08-1.53), uso de fármacos para gastritis: RP 1.24 (IC 95% 1.07-1.42), diabetes de más de 10 años de duración: RP 1.35 (IC 95% 1.08-1.42), más de dos tabletas de hipoglicemiantes al día: RP 1.29 (IC 1.12-1.48), y uso de insulina: RP 1.29 (IC 95% 1.12-1.47). Las variables o factores que se incluyeron en el modelo multivariable final fueron: uso inadecuado de fármacos, uso fármacos para la gastritis, uso de insulina, diabetes de más de diez años de duración y no cumplimiento de la dieta. Conclusiones: se encontró una alta proporción de pacientes diabéticos con falta de control metabólico, asociado a factores relacionados con la enfermedad como duración de la diabetes, fármacos, dieta e insulina. Otros factores de carácter personal fueron: bajo apoyo familiar, falta de soporte económico y estar desempleado. (Acta Med Colomb 2013; 38: 213-221).

Background: the prevalence of failure in the metabolic control in diabetic patients and the risk factors associated with it in the city of Bucaramanga are not known. Material and Methods: cross- sectional analytical study which measured the prevalence of no metabolic control in type 2 diabetic patients, from the promotion and prevention of cardiovascular risk program in a Health Promoting Company in Bucaramanga. The no metabolic control was measured by glycosylated hemoglobin (greater than 7.0 %). the effects of the assessed variables (personal, clinical and perception of care) were estimated through prevalence ratios (PR) and a multivariate model. Results: 484 patients were included, of whom 407 had Hba1c record during the 3 months preceding the survey. the prevalence of no metabolic control was 65%. the variables associated with this outcome with p < 0.05 were: history of dyslipidemia: Pr 1.14 (Ci 1.00-1.32), non-adherence to medication: Pr 1.6 (Ci 1.01-1.34), having between 50 and 65 years: RP 1.16 (Ci 1.01-1.34), consuming more than six medication tablets daily : RP 1.22 (Ci 1.06-1.41), being unemployed: PR 1.38 (CI 1.09-1.76), assist sometimes accompanied to the medical consultation: PR 1.29 (CI 1.08-1.53), use of drugs for gastritis: Pr 1.24 (95% CI 1.07-1.42), more than 10 years' duration of diabetes: PR 1.35 (95% 1.08-1.42), more than two tablets daily of hypoglycemic drugs: PR 1.29 (CI 1.12-1.48), and use of insulin: PR 1.29 (95% CI 1.12-1.47). The variables or factors that were included in the final multivariate model were: inappropriate use of drugs, use of drugs for gastritis, insulin use, duration of diabetes for more than ten years and non-dietary compliance. Conclusions: we found a high proportion of diabetic patients with poor metabolic control associated with disease-related factors such as duration of diabetes, drugs, diet and insulin. Other personal factors were: low family support, lack of financial support and being unemployed. (Acta Med Colomb 2013; 38: 213-221).

Maternal height and child growth patterns.

OBJECTIVE: To examine associations between maternal height and child growth during 4 developmental periods: intrauterine, birth to age 2 years, age 2 years to mid-childhood (MC), and MC to adulthood. STUDY DESIGN: Pooled analysis of maternal height and offspring growth using 7630 mother-child pairs from 5 birth cohorts (Brazil, Guatemala, India, the Philippines, and South Africa). We used conditional height measures that control for collinearity in height across periods. We estimated associations between maternal height and offspring growth using multivariate regression models adjusted for household income, child sex, birth order, and study site. RESULTS: Maternal height was associated with birth weight and with both height and conditional height at each age examined. The strongest associations with conditional heights were for adulthood and 2 years of age. A 1-cm increase in maternal height predicted a 0.024 (95% CI: 0.021-0.028) SD increase in offspring birth weight, a 0.037 (95% CI: 0.033-0.040) SD increase in conditional height at 2 years, a 0.025 (95% CI: 0.021-0.029 SD increase in conditional height in MC, and a 0.044 (95% CI: 0.040-0.048) SD increase in conditional height in adulthood. Short mothers (<150.1 cm) were more likely to have a child who was stunted at 2 years (prevalence ratio = 3.20 (95% CI: 2.80-3.60) and as an adult (prevalence ratio = 4.74, (95% CI: 4.13-5.44). There was no evidence of heterogeneity by site or sex. CONCLUSION: Maternal height influences offspring linear growth over the growing period. These influences likely include genetic and non-genetic factors, including nutrition-related intergenerational influences on growth that prevent the attainment of genetic height potential in low- and middle-income countries.

Fatores associados à anemia em lactentes nascidos a termo e sem baixo peso / Anemia-associated factors in infants born at term with normal weight

OBJETIVO: Investigar os fatores envolvidos na gênese da anemia ferropriva em lactentes. MÉTODOS: Trata-se de um estudo transversal, no qual foram avaliadas 104 crianças no segundo ano de vida, que nasceram a termo e sem baixo peso, no município de Viçosa, Minas Gerais. Foi aplicada entrevista aos pais, realizado recordatório 24 horas e avaliação antropométrica. Os exames laboratoriais foram eritrograma, ferritina e retinol sérico. O estudo foi aprovado pelos Comitês de Ética com seres humanos da UFMG e UFV. As análises estatísticas foram conduzidas no Epi Info e SPSS. O Modelo de Regressão Linear generalizado de Poisson, com resultados expressos como razões de prevalências, foi utilizado para verificar a associação da anemia com as variáveis do estudo. RESULTADOS: A deficiência de vitamina A e a anemia estiveram presentes em 9,6 por cento e 26 por cento das crianças, respectivamente. A anemia dos lactentes se associou ao uso de composto ferroso no pós-parto pela mãe, uso anterior de composto ferroso pela criança, início do pré-natal, tempo de aleitamento materno predominante e condição de trabalho do pai. Assim, no segundo ano de vida, os lactentes filhos de mulheres que iniciaram o pré-natal tardiamente e não usaram o composto ferroso após o parto, com pais em situação de desemprego, que nunca receberam composto ferroso e que mantiveram o aleitamento materno predominante por mais de quatro meses, apresentaram significantemente maior prevalência de anemia. CONCLUSÃO: Os resultados demonstraram a importância da nutrição durante a gestação e durante a infância na prevenção da anemia em crianças.

OBJECTIVE: To investigate the factors involved in the genesis of infant iron deficiency anemia. METHODS: This is a cross-sectional study, which evaluated 104 children in their second year of life who were born at term with adequate weight in Viçosa, Minas Gerais, Brazil. An interview, a 24-hour recall to parents, and anthropometric assessment were used. Laboratory exams included blood count, ferritin, and serum retinol. This study was approved by the Ethics Committee of UFMG and UFV. Statistical analyses were conducted using the Epi Info and SPSS softwares. Poisson generalized linear regression model was used to determine the association of anemia with the study variables, with results expressed as prevalence ratio. RESULTS: Vitamin A deficiency and anemia were identified in 9.6 percent e 26 percent of the children, respectively. Infant anemia was associated with the date of onset of prenatal care, maternal use of iron after childbirth, paternal working status, prior use of iron by the child, and duration of breastfeeding. Thus, in the second year of life, lactating children of women who began prenatal care late and did not use iron compounds after birth, with unemployed parents, who never received iron compounds, and who were predominantly breastfed for more than four months had significantly higher prevalence of anemia. CONCLUSION: The results have demonstrated the importance of nutrition during pregnancy and infancy in the prevention of anemia in children.

Risk factors for oral candidiasis in Brazilian HIV-infected adult patients

Aim: The goals of this study were: 1) to estimate the prevalence of oral candidiasis (OC) in a sample of Brazilian HIV-infected adult patients, and 2) to investigate the risk factors for HIVassociated OC in this sample. Methods: This case-control study included 112 HIV-infected patients treated between 2002 and 2004 at a clinic for sexually transmitted diseases. Data were collected from medical records and clinical examinations. Diagnosis of OC was performed in accordance with the International Classification System and cytological features. Seventeen clinical and laboratorial variables were registered. Univariate analyses were performed on all variables. Multiple logistic regression techniques were used to develop a model and identify the set of variables that may predict risk factors in HIV-infected adult patients with OC. Results: Prevalence of OC was 31.3%. OC was associated with oral hairy leukoplakia (OHL) [p<0.001; odds ratio (OR) = 10.2 (95%CI: 4.0-26.0)], previous use of fluconazole [p<0.001; OR=27.4 (95%CI: 8.1-92.0)] and viral load [p=0.042; OR=2.3 (95%CI: 1.0-5.3)]. Conclusions: These results are important for the development of strategies to eliminate these risk factors and significantly reduce OC in HIV-infected patients.

Métodos para estimar razão de prevalência em estudos de corte transversal / Métodos para estimar razón de prevalencia en estudios de cohorte transversal / Methods for estimating prevalence ratios in cross-sectional studies

OBJETIVO: Comparar empiricamente as regressões de Cox, log-binomial, Poisson e logística para estimar razões de prevalência em estudos de corte transversal. MÉTODOS: Foram utilizados dados de um estudo epidemiológico transversal (n=2.072), de base populacional, realizado com idosos na cidade de São Paulo (SP), entre maio de 2003 e abril de 2005. Diagnósticos de demência, possíveis casos de transtorno mental comum e autopercepção de saúde ruim foram escolhidos como desfechos com prevalência baixa, intermediária e alta, respectivamente. Foram utilizadas variáveis de confusão com duas ou mais categorias ou valores contínuos. Valores de referência para estimativas por ponto e por intervalo para as razões de prevalência (RP) foram obtidos pelo método de estratificação de Mantel-Haenszel. Estimativas ajustadas foram calculadas utilizando regressões de Cox e Poisson com variância robusta, e regressão log-binomial. Odds ratios (OR) brutos e ajustados foram obtidos pela regressão logística. RESULTADOS: As estimativas por ponto e por intervalo obtidas pelas regressões de Cox e Poisson foram semelhantes à obtida pela estratificação de Mantel-Haenszel, independentemente da prevalência do desfecho e das covariáveis do modelo. O modelo log-binomial apresentou dificuldade de convergência quando o desfecho tinha prevalência alta e havia covariável contínua no modelo. A regressão logística produziu estimativas por ponto e por intervalo maiores do que as obtidas pelos outros métodos, principalmente para os desfechos com maiores prevalências iniciais. Se interpretados como estimativas de RP, os OR superestimariam as associações para os desfechos com prevalência inicial baixa, intermediária e alta em 13 por cento, quase 100 por cento e quatro vezes mais, respectivamente. CONCLUSÕES: Em análise de dados de estudos de corte transversal, os modelos de Cox e Poisson com variância robusta são melhores alternativas que a regressão logística. O modelo de regressão log-binomial produz e...