Principal components and generalized linear modeling in the correlation between hospital admissions and air pollution

OBJECTIVE To analyze the association between concentrations of air pollutants and admissions for respiratory causes in children. METHODS Ecological time series study. Daily figures for hospital admissions of children aged < 6, and daily concentrations of air pollutants (PM10, SO2, NO2, O3 and CO) were analyzed in the Região da Grande Vitória, ES, Southeastern Brazil, from January 2005 to December 2010. For statistical analysis, two techniques were combined: Poisson regression with generalized additive models and principal model component analysis. Those analysis techniques complemented each other and provided more significant estimates in the estimation of relative risk. The models were adjusted for temporal trend, seasonality, day of the week, meteorological factors and autocorrelation. In the final adjustment of the model, it was necessary to include models of the Autoregressive Moving Average Models (p, q) type in the residuals in order to eliminate the autocorrelation structures present in the components. RESULTS For every 10:49 μg/m3 increase (interquartile range) in levels of the pollutant PM10 there was a 3.0% increase in the relative risk estimated using the generalized additive model analysis of main components-seasonal autoregressive – while in the usual generalized additive model, the estimate was 2.0%. CONCLUSIONS Compared to the usual generalized additive model, in general, the proposed aspect of generalized additive model − principal component analysis, showed better results in estimating relative risk and quality of fit.

Determinants of using pacifier and bottle feeding

OBJECTIVE To analyze the factors associated with the use of pacifiers and/or bottle feeding in infants aged under one year.METHODS This is a cross-sectional study with 34,366 children and using data from the database of the 2nd Nationwide Survey of Breastfeeding Prevalence performed in the Brazilian capitals and Federal District in 2008. Cluster sampling was used. The questionnaire included questions about the use of artificial nipples in the last 24 hours. The analysis considered three outcomes: exclusive use of pacifier, exclusive use of bottle feeding, and use of artificial nipples (pacifier and bottle feeding). Prevalence ratios were obtained using Poisson regression with robust variance following a hierarchical model.RESULTS The following factors were associated with exclusive use of the pacifier: mother working outside the home, primiparity, child was not breastfed within the first hour, and child had consumed tea on the first day at home. The following factors were associated with exclusive use of bottle feeding: mother working outside the home, primiparity, low birth weight, child not breastfed within the first hour, and child had consumed milk formula and tea on the first day at home. The following factors were associated with use of artificial nipples (pacifier and bottle feeding): mother working outside the home, primiparity, cesarean delivery, the male gender, low birth weight, born in a hospital not accredited as “baby friendly”, required health baby monitoring in the Primary Health Care Unit (PR = 0.91), and child had consumed milk formula, water, or tea on the first day at home.CONCLUSIONS This study identified profiles of exclusive users of pacifiers, bottle feeding, and both. The provided information can guide preventive practices for child health.

Temporal evolution and spatial distribution of maternal death

OBJECTIVE To analyze the temporal evolution of maternal mortality and its spatial distribution.METHODS Ecological study with a sample made up of 845 maternal deaths in women between 10 and 49 years, registered from 1999 to 2008 in the state of Rio Grande do Sul, Southern Brazil. Data were obtained from Information System on Mortality of Ministry of Health. The maternal mortality ratio and the specific maternal mortality ratio were calculated from records, and analyzed by the Poisson regression model. In the spatial distribution, three maps of the state were built with the rates in the geographical macro-regions, in 1999, 2003, and 2008.RESULTS There was an increase of 2.0% in the period of ten years (95%CI 1.00;1.04; p = 0.01), with no significant change in the magnitude of the maternal mortality ratio. The Serra macro-region presented the highest maternal mortality ratio (1.15, 95%CI 1.08;1.21; p < 0.001). Most deaths in Rio Grande do Sul were of white women over 40 years, with a lower level of education. The time of delivery/abortion and postpartum are times of increased maternal risk, with a greater negative impact of direct causes such as hypertension and bleeding.CONCLUSIONS The lack of improvement in maternal mortality ratio indicates that public policies had no impact on women’s reproductive and maternal health. It is needed to qualify the attention to women’s health, especially in the prenatal period, seeking to identify and prevent risk factors, as a strategy of reducing maternal death.

Dimensions of self-rated health in older adults

OBJECTIVE To analyze the association between negative self-rated health and indicators of health, wellbeing and sociodemographic variables in older adults. METHODS Cross-sectional study that used data from a population-based health survey with a probability cluster sample that was carried out in Campinas, SP, Southeastern Brazil,, in 2008 and 2009. The participants were older adults (≥ 60 years) and the dependent variable was self-rated health, categorized as: excellent, very good, good, bad and very bad. The adjusted prevalence ratios were estimated by means of Poisson multiple regression. RESULTS The highest prevalences of bad/very bad self-rated health were observed in the individuals who never attended school, in those with lower level of schooling, with monthly per capita family income lower than one minimum salary. Individuals who scored five or more in the physical health indicator also had bad self-rated health, as well as those who scored five or more in the Self-Reporting Questionnaire 20 and those who did not refer feeling happiness all the time. CONCLUSIONS The independent effects of material life conditions, physical and mental health and subjective wellbeing, observed in self-rated health, suggest that older adults can benefit by health policies supported by a global and integrative view of old age.

Back pain in adults living in quilombola territories of Bahia, Northeastern Brazil

OBJECTIVE To analyze the factors associated with back pain in adults who live in quilombola territories. METHODS A population-based survey was performed on quilombola communities of Vitória da Conquista, state of Bahia, Northeastern Brazil. The sample (n = 750) was established via a raffle of residences. Semi-structured interviews were conducted to investigate sociodemographics and employment characteristics, lifestyle, and health conditions. The outcome was analyzed as a dichotomous variable (Poisson regression). RESULTS The prevalence of back pain was of 39.3%. Age ≥ 30 years and being a smoker were associated with the outcome. The employment status was not related to back pain. CONCLUSIONS The survey identified a high prevalence of back pain in adults. It is suggested to support the restructuring of the local public service in order to outline programs and access to healthy practices, assistance, diagnosis, and treatment of spine problems.

Women with HIV: gender violence and suicidal ideation

OBJECTIVE To analyze the relationship between gender violence and suicidal ideation in women with HIV. METHODS A cross-sectional study with 161 users of specialized HIV/AIDS care services. The study investigated the presence of gender violence through the Brazilian version of the World Health Organization Violence against Women instrument, and suicidal ideation through the Suicidal Ideation Questionnaire. Statistical analyses were performed with the SPSS software, using the Chi-square test and Poisson multiple regression model. RESULTS Eighty-two women with HIV reported suicidal ideation (50.0%), 78 (95.0%) of who had suffered gender violence. Age at first sexual intercourse < 15 years old, high number of children, poverty, living with HIV for long, and presence of violence were statistically associated with suicidal ideation. Women who suffered gender violence showed 5.7 times more risk of manifesting suicidal ideation. CONCLUSIONS Women with HIV showed a high prevalence to gender violence and suicidal ideation. Understanding the relationship between these two grievances may contribute to the comprehensive care of these women and implementation of actions to prevent violence and suicide.

Trends in the use of antidepressants among older adults: Bambuí Project

OBJECTIVE To analyze the trends and factors associated with the antidepressant use among older adults. METHODS This population-based study evaluated older adults in 1997 (n = 351, baseline) and the survivors at the 15th follow-up year (n = 462, in 2012) among the aging cohort of Bambuí. The prevalence of antidepressant use was estimated, and the most commonly used antidepressants each year were identified. Prevalence ratios with 95% confidence intervals were estimated using Poisson regression with robust variance to investigate differences in the prevalence of use between 1997 and 2012. RESULTS The overall consumption of antidepressants (PR = 2.87, 95%CI 1.94;4.25) and of selective serotonin reuptake inhibitors (PR = 7.50, 95%CI 3.74;15.02) was significantly higher in 2012. However, no significant difference was observed in the use of tricyclic antidepressants between the two cohorts (PR = 0.89, 95%CI 0.49;1.62). In the 2012 cohort, antidepressant use was associated with females, increased age, increased income (≥ 4 minimum wages), self-assessment of health as reasonable, and attending ≥ 5 medical consultations in the last 12 months. CONCLUSIONS The increased consumption of antidepressants in the period due to increased use of selective serotonin reuptake inhibitors was consistent with results observed in international studies of different population groups and contexts. The positive correlation observed between antidepressant use and family income may be a warning of possible inequalities in access to mental health services.

Isolated and synergistic effects of PM10 and average temperature on cardiovascular and respiratory mortality

OBJECTIVE To analyze the effect of air pollution and temperature on mortality due to cardiovascular and respiratory diseases. METHODS We evaluated the isolated and synergistic effects of temperature and particulate matter with aerodynamic diameter < 10 µm (PM10) on the mortality of individuals > 40 years old due to cardiovascular disease and that of individuals > 60 years old due to respiratory diseases in Sao Paulo, SP, Southeastern Brazil, between 1998 and 2008. Three methodologies were used to evaluate the isolated association: time-series analysis using Poisson regression model, bidirectional case-crossover analysis matched by period, and case-crossover analysis matched by the confounding factor, i.e., average temperature or pollutant concentration. The graphical representation of the response surface, generated by the interaction term between these factors added to the Poisson regression model, was interpreted to evaluate the synergistic effect of the risk factors. RESULTS No differences were observed between the results of the case-crossover and time-series analyses. The percentage change in the relative risk of cardiovascular and respiratory mortality was 0.85% (0.45;1.25) and 1.60% (0.74;2.46), respectively, due to an increase of 10 μg/m3 in the PM10 concentration. The pattern of correlation of the temperature with cardiovascular mortality was U-shaped and that with respiratory mortality was J-shaped, indicating an increased relative risk at high temperatures. The values for the interaction term indicated a higher relative risk for cardiovascular and respiratory mortalities at low temperatures and high temperatures, respectively, when the pollution levels reached approximately 60 μg/m3. CONCLUSIONS The positive association standardized in the Poisson regression model for pollutant concentration is not confounded by temperature, and the effect of temperature is not confounded by the pollutant levels in the time-series analysis. The simultaneous exposure to different levels of environmental factors can create synergistic effects that are as disturbing as those caused by extreme concentrations.

Association between food assistance program participation and overweight

OBJECTIVE The objective of this study was to investigate the association between food assistance program participation and overweight/obesity according to poverty level. METHODS A cross-sectional analysis of data from 46,217 non-pregnant and non-lactating women in Lima, Peru was conducted; these data were obtained from nationally representative surveys from the years 2003, 2004, 2006, and 2008-2010. The dependent variable was overweight/obesity, and the independent variable was food assistance program participation. Poisson regression was used to stratify the data by family socioeconomic level, area of residence (Lima versus the rest of the country; urban versus rural), and survey year (2003-2006 versus 2008-2010). The models were adjusted for age, education level, urbanization, and survey year. RESULTS Food assistance program participation was associated with an increased risk of overweight/obesity in women living in homes without poverty indicators [prevalence ratio (PR) = 1.29; 95% confidence interval (CI) 1.06;1.57]. When stratified by area of residence, similar associations were observed for women living in Lima and urban areas; no associations were found between food assistance program participation and overweight/obesity among women living outside of Lima or in rural areas, regardless of the poverty status. CONCLUSIONS Food assistance program participation was associated with overweight/obesity in non-poor women. Additional studies are required in countries facing both aspects of malnutrition.

Socioeconomic inequalities in the access to and quality of health care services

OBJECTIVE To assess the inequalities in access, utilization, and quality of health care services according to the socioeconomic status. METHODS This population-based cross-sectional study evaluated 2,927 individuals aged ≥ 20 years living in Pelotas, RS, Southern Brazil, in 2012. The associations between socioeconomic indicators and the following outcomes were evaluated: lack of access to health services, utilization of services, waiting period (in days) for assistance, and waiting time (in hours) in lines. We used Poisson regression for the crude and adjusted analyses. RESULTS The lack of access to health services was reported by 6.5% of the individuals who sought health care. The prevalence of use of health care services in the 30 days prior to the interview was 29.3%. Of these, 26.4% waited five days or more to receive care and 32.1% waited at least an hour in lines. Approximately 50.0% of the health care services were funded through the Unified Health System. The use of health care services was similar across socioeconomic groups. The lack of access to health care services and waiting time in lines were higher among individuals of lower economic status, even after adjusting for health care needs. The waiting period to receive care was higher among those with higher socioeconomic status. CONCLUSIONS Although no differences were observed in the use of health care services across socioeconomic groups, inequalities were evident in the access to and quality of these services.

Violence against metropolitan bus drivers and fare collectors in Brazil

OBJECTIVE To analyze the correlation between sociodemographic factors and working conditions of bus workers in a metropolitan area and violence against them.METHODS This cross-sectional study used a nonprobabilistic sample estimated according to the number of workers employed in bus companies located in three cities in the Belo Horizonte metropolitan region in 2012 (N = 17,470). Face-to-face interviews were conducted using a digital questionnaire. The factors associated with violence were analyzed in two stages using Poisson regression, according to each level. The magnitude of the association was evaluated using prevalence ratios with robust variance and a statistical significance of 5%, and 95% confidence intervals were obtained.RESULTS The study sample comprised 782 drivers and 691 fare collectors; 45.0% participants reported at least one act of violence in the workplace in the last 12 months, with passengers being predominantly responsible. The age of the bus workers was inversely associated with violence. Chronic diseases, sickness absenteeism, and working conditions were also associated with violence.CONCLUSIONS The findings on the correlation between violence and working conditions are essential for implementing prevention strategies by transportation service managers.

Modeling for control design of a multi-reach water canal

This article addresses the problem of obtaining reduced complexity models of multi-reach water delivery canals that are suitable for robust and linear parameter varying (LPV) control design. In the first stage, by applying a method known from the literature, a finite dimensional rational transfer function of a priori defined order is obtained for each canal reach by linearizing the Saint-Venant equations. Then, by using block diagrams algebra, these different models are combined with linearized gate models in order to obtain the overall canal model. In what concerns the control design objectives, this approach has the advantages of providing a model with prescribed order and to quantify the high frequency uncertainty due to model approximation. A case study with a 3-reach canal is presented, and the resulting model is compared with experimental data. © 2014 IEEE.

Estimation and hypothesis testing in mixed linear models

Dissertação apresentada para obtenção do Grau de Doutor em Matemática, Estatística, pela Universidade Nova de Lisboa, faculdade de Ciências e Tecnologia

Incidence and risk factors for intimate partner violence during the postpartum period

OBJECTIVE To estimate the incidence and identify risk factors for intimate partner violence during postpartum.METHODS This prospective cohort study was conducted with women, aged between 18-49 years, enrolled in the Brazilian Family Health Strategy in Recife, Northeastern Brazil, between 2005 and 2006. Of the 1.057 women interviewed during pregnancy and postpartum, 539 women, who did not report violence before or during pregnancy, were evaluated. A theoretical-conceptual framework was built with three levels of factors hierarchically ordered: women’s and partners’ sociodemografic and behavioral characteristics, and relationship dynamics. Incidence and risk factors of intimate partner violence were estimated by Poisson Regression.RESULTS The incidence of violence during postpartum was 9.3% (95%CI 7.0;12.0). Isolated psychological violence was the most common (4.3%; 95%CI 2.8;6.4). The overlapping of psychological with physical violence occurred at 3.3% (95%CI 2.0;5.3) and with physical and/or sexual in almost 2.0% (95%CI 0.8;3.0) of cases. The risk of partner violence during postpartum was increased for women with a low level of education (RR = 2.6; 95%CI 1.3;5.4), without own income (RR = 1.7; 95%CI 1.0;2.9) and those who perpetrated physical violence against their partner without being assaulted first (RR = 2.0; 95%CI 1.2;3.4), had a very controlling partner (RR = 2.5; 95%CI 1.1;5.8), and had frequent fights with their partner (RR = 1.7; 95%CI 1.0;2.9).CONCLUSIONS The high incidence of intimate partner violence during postpartum and its association with aspects of the relationship’s quality between the couple, demonstrated the need for public policies that promote conflict mediation and enable forms of empowerment for women to address the cycle of violence.

Bullying among adolescents in a Brazilian urban center – “Health in Beagá” Study

OBJECTIVE To analyze the prevalence of bullying and its associated factors in Brazilian adolescents.METHODS Data were used from a population-based household survey conducted by the Urban Health Observatory (OSUBH) utilizing probability sampling in three stages: census tracts, residences, and individuals. The survey included 598 adolescents (14-17 years old) who responded questions on bullying, sociodemographic characteristics, health-risk behaviors, educational well-being, family structure, physical activity, markers of nutritional habits, and subjective well-being (body image, personal satisfaction, and satisfaction with their present and future life). Univariate and multivariate analysis was done using robust Poisson regression.RESULTS The prevalence of bullying was 26.2% (28.0% among males, 24.0% among females). The location of most bullying cases was at or on route to school (70.5%), followed by on the streets (28.5%), at home (9.8%), while practicing sports (7.3%), at parties (4.6%), at work (1.7%), and at other locations (1.6%). Reports of bullying were associated with life dissatisfaction, difficulty relating to parents, involvement in fights with peers and insecurity in the neighborhood.CONCLUSIONS A high prevalence of bullying among participating adolescents was found, and the school serves as the main bullying location, although other sites such as home, parties and workplace were also reported. Characteristics regarding self-perception and adolescent perceptions of their environment were also associated with bullying, thus advancing the knowledge of this type of violence, especially in urban centers of developing countries.