A STINT Institutional Grants ProgrammeSocial determinants of health - Swedish-Brazilian exchange

THE 1st IMS/CHESS WORKSHOP: Health and Society8 – 10 December 2008

Centro de visitantes, Floresta da TijucaRio de Janeiro, Brazil

WELCOME NOTE

WELCOME to the 1st IMS/CHESS Workshop whose core theme is “Health and Society”.

This is only the first of a (hopefully!) long series of workshops that will take place under the

exchange program between IMS/UERJ and CHESS. This exchange program is funded by the

Swedish Foundation for International Cooperation in Research and Higher Education (STINT)

that is set to last for four years, starting in 2007.

We do hope all participants enjoy the programme of talks and oral presentations given by

Swedish and Brazilian Researchers as well as Doctorate Students. When planning the scientific

programme our goal was to encourage debates and exchange of ideas, thus allowing every

participant of the workshop to learn and reflect upon relevant research topics being developed

both at the IMS and at CHESS.

Our choice of venue was made according to what is most valued, in our opinion, by Brazilian

and Swedish people: nature at its best and impressive shape. So we also hope you appreciate

the surroundings and inspire yourself with new insights and new breakthroughs in your work.

The social programme is simple, but very exciting! First we invite all to dinner1 at the old and

traditional restaurant located at the “Pracinha do Alto”, at the entrance of “Floresta da Tijuca”,

on Tuesday evening after the scientific activities. Then on Wednesday afternoon, everybody is

invited to walk to the very top of Rio de Janeiro city, in a guided excursion, so as to minimize

the risk of getting lost in this amazing forest, located at the heart of a city of 7 million

inhabitants.

So prepare yourself physically and mentally for a very special experience of natural and

scientific endurance during the three days of the IMS-CHESS, Rio-Stockholm, Brazil-Sweden

workshop.

Thanks for your participation,

Johan Fritzell Antonio Ponce de Leon

1 Unfortunately, expenses will not be covered by the workshop organization.

2

3

PROGRAMME

Monday December 8th 0900 – Arrive at venue 0930 – 1015 – Welcome and introduction of STINT collaboration (Antonio Ponce de Leon and Johan Fritzell) 1015 – 1115 – Lecture: Johan Fritzell : “The Luxembourg Income Study – Past, present and future” 1115 – 1230 – Brunch 1230 – 1330 – Lecture: Eduardo Faerstein: “Race” and Health: Some Trends and Challenges 1330 – 1430 – Lecture: Antonio Ponce de Leon: “Health Resource Allocation in the Stockholm County” 1430 – 1500 – Coffee/tea 1500 – 1630 – Panel Round table discussion, Chair: Antonio Ponce de Leon. Kenneth Rochel de Camargo (to be confirmed), Alba Maria Zaluar, Maria Kölegård and Johan Fritzell discuss around Closing the gap in a generation. Health equity through action on the social determinants of health, the Final report of the Commission on social determinants of health, http://www.who.int/social_determinants/final_report/en/index.html 1630 - 1645– Closing of the first day

4

Tuesday December 9th

Presentations are 40 minutes (20 min talk/20 min comments + discussion) 0900 – 1100 – Social determinants of health (Chair: Luciana Tricai Cavalini) Presentation 1: Roger Keller Celeste: “Income inequality and lag time: does the data fit the theory?” Comments by Johan Fritzell Presentation 2: Carlos Castro Luz: “Healthcare indicators related to pre-hospital care policies for the elderly in areas in Southeast Brazil” Comments by Mikael Rostila Presentation 3: Roger Keller Celeste: “ Trends in socioeconomic disparities in oral health in Brazil and Sweden” Comments by Antonio Ponce de Leon 1100 – 1215 – Brunch 1215 – 1415 – Work and health (Chair: Antonio Ponce de Leon) Presentation 4: Odaleia B Aguiar: “Reproducibility of version “Demand-Control Questionnaire (DCQ)” among industrial restaurant workers of Rio de Janeiro State” Comments by Susanna Toivanen Presentation 5: Yara H M Hökerberg: “Validation study of the Brazilian version of Demand Control Support Questionnaire” Comments by Washington Leite Junger Presentation 6: Susanna Toivanen: “Psychosocial work environment and risk factors for CVD; cross sectional and change over time in workers in Sweden” Comments by Cláudia de Souza Lopes 1415 – 1445 – Coffee/tea 1445 – 1645 – Violence (Chair: Maria Kölegård) Presentation 7: Washington Leite Junger: “Urban violence: spatial dynamics, people perception, and socioeconomic relations in Rio de Janeiro” Comments by Johan Fritzell Presentation 8: Alessandra Dias Oliveira: “The intimate partner violence during pregnancy: Is it a risk factor to premature interruption of exclusive breastfeeding duration? “ Comments by Michael Reichenheim Presentation 9: Gustavo Lobato: “Intimate partner violence during the pregnancy: a risk factor to postpartum depressive symptoms?” Comments by Luciana Tricai Cavalini 1700 – Get together and dinner at the Pracinha restaurant, Alto da Boa Vista

5

Wednesday December 10th Presentations are 40 minutes (20 min talk/20 min comments + discussion) 0900 – 1020 – Social determinants of health (Chair: Johan Fritzell) Presentation 10: Jacyara Maria Brito Macedo “HPV infection and associated factors among women from low income area of Rio de Janeiro city” Comments by Eduardo Faerstein Presentation 11: Ilana Nogueira Bezerra: “Eating out of home and obesity: a Brazilian nationwide survey” Comments by Maria Kölegård 1020 – 1030 – Break 1030 – 1230 – Ethnicity/race and health (Chair: Susanna Toivanen) Presentation 12: Karine de Lima Sírio Boclin:“Racial/ethnic and socioeconomic inequalities in occurrence of uterine myomas” Comments by Johan Fritzell Presentation 13: João Luiz Dornelles Bastos: “Racial inequalities in health: measuring self-reported experiences of discrimination in Brazil” Comments by Rosely Sichieri Presentation 14: Mikael Rostila: ” Birds of a feather flock together – and fall ill? Migrant homophily and health in Sweden” Comments by Eduardo Faerstein 1230 – 1330 – Lunch 1330 – 1400 – Closing the workshop and looking ahead. Antonio Ponce de Leon and Johan Fritzell 1400 - 1645 – Guided hike in the national park Floresta da Tijuca

6

ABSTRACTS

Lectures Monday, Dec 8th Johan Fritzell: The Luxembourg Income Study – Past, present and future The Luxembourg Income Study (LIS) is regarded as the most comparable data source for cross-national studies on poverty and distribution of income. In this lecture I will first give an overview of the historical background of LIS and secondly give some empirical facts and lessons learned from LIS. LIS-data was recently used by the high level WHO Commission on Social Determinants of Health, and illustrations will be given to why LIS-data could, and should, be of increasing importance to research on social determinants of health. For the present Sweden-Brazil collaboration I will especially highlight an on-going expansion of LIS to include more middle-income countries, among them Brazil.

7

Presentations Tuesday, Dec 9th

Roger Keller Celeste: Income Inequality and Lag Time: Does The Data Fit the Theory? Roger Keller CELESTE, Johan FRITZELL, Paulo NADANOVSKY Affiliations: Department of Epidemiology (IMS-UERJ), Centre for Health Equity Studies, CHESS (Stockholm University –Karolinska Institute) To evaluate the association between income inequality at lagged time of 2 and 11 years and two short latency outcomes (untreated dental caries and gingivitis) and two long latency outcomes (edentulism and periodontal attachment loss >8mm). Methods: We used data from an oral health survey in Brazil in 2002-2003. Our analysis included 22,838 35-44 year-olds subjects clustered in 330 municipalities. Different lagged Gini at municipal level, from 2000 and 1991, were fitted using logistic and negative binomial multilevel analysis. Individual level covariates included equivalised income, age, sex, time since last dental visit and place of residence (rural versus urban). Municipal per capita income was also included in adjusted models. Comparisons among models were done based on Akaike’s Information Criteria. Results: Crude estimates showed that only untreated dental caries was associated with current and lagged Gini of 1991. Adjusted models showed that only Gini in 2000 remained statistically significant, rate ratio of 1.19 (95% CI 1.09-1.30). Overall, alternative models with no Gini coefficient yielded best fit. The proportion of the total variance attributable to municipal level was considerable in all models (lowest 9.5% and highest 31.8%). Conclusions: Income inequality was not cumulatively associated with long term outcomes. Regarding shorter latency outcomes, it was associated with untreated dental caries but not with periodontal diseases. The fact that dental caries is more prone to prevention through primary care and public policies may explain the association with the former but not with the later.

8

Carlos Castro Luz : Healthcare indicators related to pre-hospital care policies for the elderly in areas in Southeast Brazil

Carlos Castro Luz, Luciana Tricai Cavalini

The quick increase of the elderly population is becoming a global phenomenon. Several studies have tried to analyze the possible causes and consequences of this process. In Brazil, in parallel with the demographic transition, an epidemiologic transition has occurred, with increase of the non-transmittable diseases. This transition trends to change the morbidity and mortality patterns in the population and it is associated with other socioeconomic transformations. The objective of this paper is to draw a profile of the morbidity and the mortality of population aged 60 years and more in three regions of the wealthiest region in Brazil: the first with a metropolitan area, the second with a big size city and the third with a medium size city for the year 2005. Data were collected from the Brazilian public demographic and health information systems freely available on internet and analyzed with the open source statistical package R. For all regions, the main causes of hospital admissions and deaths for the elderly population were: cardiovascular diseases, neoplasms, respiratory diseases and ill-defined causes of morbidity and mortality. Despite the different size of the municipalities, all regions presented a very similar profile in terms of morbidity and mortality. This situation configures a serious problem for healthcare management of small towns, since their healthcare budget is smaller, their demand for qualified healthcare professionals is less covered, and their access to higher complexity levels of healthcare is more limited. We conclude that the public health policies for the elderly population should take into account the lack of resources of smaller municipalities in Brazil, since there is a lack of preparedness, in terms of healthcare infrastructure, for them to face the challenges of the ageing process of their population.

9

Roger Keller Celeste: Trends in socioeconomic disparities in oral health in Brazil and Sweden URoger Keller CELESTE, Johan FRITZELL, Paulo NADANOVSKY Affiliations: Department of Epidemiology (IMS-UERJ). Centre for Health Equity Studies, CHESS (Stockholm University –Karolinska Institute). It has been suggested that new interventions, as time goes by, may initially increase socioeconomic inequalities in health to decrease afterwards, the so called Inverse Equity Hypothesis. The dynamics of trends in inequalities is not well understood yet. Our objective was to describe trends in socioeconomic gaps in oral health in Brazil and Sweden among adults (35-44 year-old). Methods: This study is a time trend analysis of cross-sectional studies designed to assess the prevalence of selected oral health outcomes and other population characteristics. In Sweden we obtained data for the years 1968, 1974, 1981, 1991 and 2000. In Brazil data were available for 1986 and 2002. Trends in the prevalence of two outcomes were shown by socioeconomic group: a) edentulism and b) good own teeth (Sweden)/no missing tooth (Brazil). Results: In 1968, in Sweden, the prevalence difference between two socioeconomic groups in edentulism was 24.3% (95% CI= 15.9-32.8) and in teeth in good conditions was 20.8% (95% CI= 13.6-28.1). In 1986, in Brazil, the prevalence difference in edentulism was 9.2% (95% CI= 3.2-15.2) and for no missing tooth was 3.0% (95% CI= 0.8-5.3). Regarding edentulism, there was on average an annual decline in absolute disparities of 0.4% (95% CI= 0.2-0.7) in Brazil and 0.7% (95% CI= 0.5-0.9) in Sweden. Results concerning no missing teeth (Brazil) / teeth in good conditions (Sweden) were mixed. In Brazil the absolute and relative disparities in the prevalence of no missing teeth has increased, while in Sweden there was a statistically non significant decrease in absolute disparities and a small reduction in relative disparities. Since 1991 in Sweden and 2002 in Brazil, we found that, in the age group of 35-44, there are no significant socioeconomic disparities in edentulism either in absolute or relative terms. Conclusions: Despite improvements in oral health in both countries, lower socioeconomic groups have not reached the same level of good oral health as their richer counterparts. This dynamics may suggest that, when major reduction in edentulism started in the lower economic group, the majority of people in the richer group had already improved their oral health.

10

Odaleia B. Aguiar: Reproducibility of version “Demand-Control Questionnaire (DCQ)” among industrial restaurant’s workers of Rio de Janeiro State Aguiar, Odaleia B.; Valente, Joaquim, G.; Fonseca, Maria de Jesus M. Introduction: The “Demand-Control Questionnaire (DCQ)” is focused on two crucial work dimensions, psychological demands and decision latitude on the work. Psychological demands refer to workload and conflicts demands. Decision latitude refers to the ability to make decisions about one’s work and the possibility of being creative using or developing new skills. Objective: To verify if the DCQ's Version for Portuguese would be reproducible in workers' population with low education. Methods: The eligible population for the test-retest was 52 workers, distributed in three Industrial Restaurants of one Rio de Janeiro's Concessionaire Company, whom have been interviewed by a trained interviewer, in two moments, with interval from 7 to 15 days. Measures of stability used were the intraclass correlation coefficient (ICC) and the weighted statistics kappa. Bland & Altman graphs were built. For internal consistency evaluation, the Cronbach's alpha coefficient was used. Results: The results demonstrated good reliability for total score, ICC for psychological demand, control on the work and social support at work were 0,70, 0,68 and 0,80, respectively. Cronbach's alpha in the test presented the next results 0,75, 0,50 and 0,82 to the dimensions previously cited in the sequence above, in retest. The internal consistency for the decision latitude was low for food worker. Conclusions: According to these results, the instrument reliability was considered adequate, enabling its utilization studies of association between stress at work and outcomes related to health.

11

Yara H. M. Hökerberg: Validation study of the Brazilian version of Demand Control Support Questionnaire Yara H. M. Hökerberg, Odaleia B. Aguiar, Eduardo Faerstein, Joaquim Valente, Michael Reichenheim, Maria J. Fonseca e Sonia R. Lambert Passos. Introduction: The Swedish Demand-Control-Support Questionnaire (DCSQ) is a shortened and modified version of the Job Content Questionnaire that has been used in Brazil to investigate associations between job strain and health outcomes. Objective: To explore the validity of the Brazilian version of DCSQ. Methods: This study investigated 825 workers who completed the Brazilian version of DCSQ: to 399 (48%), DCSQ was self-administered at a hospital (H) and 426 (52%) were interviewed at 9 restaurants (R) in the State of Rio de Janeiro, Brazil. Confirmatory Factor Analysis using Structural Equation Models (CFA, robust weighted least squares estimator using a diagonal weight matrix - Mplus) was used to test theoretical structure of dimensionality defined a priori. Root Mean Square Error of Approximation (RMSEA < 0.06), Weighted Root Mean Square Residual (WRMR < 0.90), Comparative Fit Index (CFI > 0.90) and Tucker-Lewis index (TLI > 0.90) were used to choose the best fit model. Results: CFA supported the instrument construct in four dimensions: demands (D), skill discretion (SD), decision authority (DA) and social support at work (SSW). Standardized factor loadings were in the range of 0.47-0.81 (D), 0.08-0.69 (SD), 0.78-0.85 (DA) and 0.63-0.81 (SSW). Since the item ‘repetitive work’ of DA displayed low standardized factor loading and job strain definition does not consider SSW, potential effect modifier of the association between job strain and health outcomes, we tested an alternative model by removing them, which resulted in better fit indices (CFI=0.918, TLI=0.910, RMSEA=0.095, WRMR=0.613). Conclusion: Although results provided evidence of the validity of four dimensions of the Brazilian DCSQ, the model with best fit was the one without SSW and the item ‘repetitive work’. Future steps will include an exploration of the validity of DCSQ applied to a similar Swedish study population using the same methods. Keywords: Validation Studies [Publication Types]; Factor Analysis; Stress, Psychological; Questionnaires.

12

Susanna Toivanen: Psychosocial work characteristics and risk factors for cardiovascular disease: cross sectional associations and change over time in workers in Sweden Susanna Toivanen & FAS project group for psychosocial factors and risk of stroke Background: Results from previous studies of the associations between job strain and major risk factors for cardiovascular disease (hypertension, hypercholesterolemia, obesity, smoking) show a great variation regarding which component of job strain (demands and decision latitude) is associated with the CVD risk factors. Results vary among genders and between different contexts. In addition, we know very little about if these associations vary over time, which could be an indication of societal influences on these associations. Aim: The aim of the present study was to: (1) analyse the associations between job strain and it components in relation to hypertension, hypercholesterolemia, obesity, and smoking, (2) explore whether these associations differ among women and men, (3) assess if the associations vary over time. Material and methods: Data came from the Northern Sweden MONICA Project, population based survey samples from the two northernmost counties in Sweden. The surveys were performed in 1986, 1990, 1994, 1999, and 2004 using stratified sampling procedures. For the present study the 1999 and 2004 surveys were analysed and all participants aged 25-64 years who reported that they worked at the time of the survey were included. The total study populations consisted of 3,320 participants in 1999 (51,2 % women) and 1,137 in 2004 (51,8 % women). Gender-specific logistic regression analyses were performed separately for the years 1999 and 2004. Demands, decision latitude, and job strain were analysed in relation to hypertension, hypercholesterolemia, waist-hip ratio, and smoking. Analyses were first adjusted for age, and in the final model for age, marital status, educational level, country of birth, self-rated health, smoking, physical inactivity, BMI and work hours. Results: In men, no significant associations were found between job strain or its components and the CVD risk factors in 1999 and 2004. Also in women, the associations were mostly weak and insignificant. However in 1999, there was an age-adjusted association between low decision latitude and smoking (OR 1,49; 95% CI 1,17-1,89). In 2004, the association between smoking and age-adjusted low decision latitude was OR 1,79 (95% CI 1,15-2,79), fully adjusted demands OR 1,65 (95% CI 1,03-2,64), and job strain OR 1,87 (95% CI 1,15-3.04). Conclusions: Demands, decision latitude and job strain were not significantly associated with hypertension, hypercholesterolemia or obesity in these women and men from Northern Sweden. In women, however, work characteristics were associated with smoking both in 1999 and 2004.

13

Washington L Junger: Urban violence: spatial dynamics, people perception, and socioeconomic relations in Rio de Janeiro The aims of this paper is to describe the spatial dynamics of urban violence, how people perceive it, and how socioeconomic factors as housing, education, and income relate to it in the city of Rio de Janeiro. We used data on deaths due to homicide from the mortality register, socioeconomic and demographic data from the latest Census and data about victimization from a household survey collected in 2005. The home addresses of the homicide victims were geographically located on a map of the city using a Geographic Information System (GIS). Smoothing techniques were applied to the point process in order to identify areas with higher concentration of victims. Cluster analysis was also used to reveal patterns on how the victims were distributed across the city. Standardised mortality ratios and empirical Bayes rates were estimated for each neighbourhood yielding results consistent with the point analysis. A factor analysis using the socioeconomic and demographic data for the census tracts was conducted and the maps of the principal components suggest that areas with poorer conditions tend to have higher concentration of people that died as a result of violence. To investigate how the people perception to violence can be explained by the socialeconomic factors we correlated responses from the victimization survey with the principal components. People perception was measured by items indicating whether people usually hear gun shots, witness some kind of misdemeanor, and whether he or she had some relative or friend killed recently. Results show that people living in the less favoured parts of the city are more likely to perceive the outcomes of urban violence.

14

Alessandra Dias Oliveira: The intimate partner violence during pregnancy: Is it a risk factor to premature interruption of exclusive breastfeeding duration?

Oliveira, ASD1; Lobato, G2; Reichenheim, ME; Moraes, CL.

Despite the recognition of Intimate Partner Violence (IPV) as a public health problem worldwide, there are still gaps in knowledge particularly regarding its potential consequence on maternal and child health. This study aimed at investigating the role of Severe Physical Violence during Pregnancy (SPVP) between partners on Exclusive Breastfeeding Duration. This is a cross-sectional study and information was collected throughout 811 interviews with mothers of children under 5 years old, who sough for health care assistance at 5 health centers in Rio de Janeiro. The information was collected between January and July of 2007 throughout face-to-face interview. To assess SPVP a national version of the Revised Conflict Tactics Scales were used. The premature breastfeeding cessation was identified by using a current status data approach, which is based on the information reported from food recall the last 7 days. The Cumulative Hazard Function was estimated by complementary log-log transformation models, which allowed the ensuing estimate of early breastfeeding cessation rates in different age groups and rates ratio of wean between women exposure and non-exposure to violence. The results suggests that SPVP is an independent risk factor of exclusive breastfeeding cessation because in the multivariate analysis women exposed to violence had an incidence-density 35% higher than those who were not exposed (RT:1,35; CI 95%: 1,01; 1,69). The finding indicates the necessity of incentives for adequately preparing health care personnel in dealing with lactating women in order to get a broader view of breastfeeding beyond the biologic aspects of lactation including the maternal psychological dimension. Particularly, it is emphasized the establishment of multi-professional teams during pre and postnatal care and awareness of the occurrence of IPV during pregnancy.

1 - Recém doutora Epidemiologia - IMS

2 - Doutorando Epidemiologia – IMS

15

Gustavo Lobato: Intimate partner violence during the pregnancy: a risk factor to postpartum depressive symptoms? Authors: Gustavo Lobato; Michael E. Reichenheim; Alessandra S. Dias; Cláudia L. Moraes. Postpartum depression (PPD) is a leading cause of maternal and child morbidity. Similarly, violence is one of the most important worldwide issues in Public Health, and several researchers have discussed that abuse in the household environment is the most common type of violence. Although several studies have already verified a consistent association between violence, physical and mental health disorders in general population, few epidemiological studies have discussed the relationship between intimate partner violence (IPV) and PPD. Thus, this research aims at evaluating if intimate partner violence during the pregnancy is a risk factor to postpartum depressive symptoms.

16

Presentations Wednesday, Dec 10th Jacyara Brito Macedo: Human papillomavirus infection and associated risk factors in a low income community of Rio de Janeiro Jacyara Maria Brito Macedo 1; Maria do Carmo Esteves da Costa 2; Mirian Carvalho de Souza 2; Liz Maria de Almeida 2; Lia Jascone da Veiga 1; Amanda Chaves Pinto 1; Gulnar Azevedo e Silva 3; 1 Instituto de Biologia Roberto Alcântara Gomes, UERJ; 2 CONPREV / INCA; 3 Instituto de Medicina Social, UERJ Introduction: The human papillomavirus (HPV) infection is the most common sexually transmitted disease (STD) and it is particularly prevalent among women. Studies using cervical samples from asymptomatic women worldwide have shown a wide range of HPV prevalence, 2 to 44 %, depending on the study area and the analysis method. The carcinogenic potential of several HPV types are very well established and the viral infection is considered a necessary although non-sufficient cause for the development of cervical cancer. Objective: The aim of this study was to estimate the prevalence of HPV DNA and to analyze the factors related to the viral infection in women from a low-income community located in the Metropolitan area of Rio de Janeiro. Methodology: A total of 1437 women aging between 18 and 59 years were identified in a census performed in the selected community of Rio de Janeiro. Of those, 662 were enrolled in this study and were attended at the health service. From a standard questionnaire it was possible to obtain information about socioeconomic status, reproductive history, sexual behavior, Papanicolaou smear history, family history of cancer and use of tobacco and alcohol. A cervical sample was collected from each woman following the conventional procedure. This material was used for cytopathology examination and for identifying HPV DNA by polymerase chain reaction (PCR). Odds ratios (crude and adjusted) were estimated using non-conditional logistic regression model. Results: The 662 interviewed and examined women represented 41.7% of the registered target population. Of those, 49.0% were less than 34 years of age and 73.0% informed to have had three partners or less in their lives. History of STD was mentioned by 8.6% of women and 82.0% informed to have performed the last Papanicolaou smear test in the last three years. Most of women were non-white (60.6%) and half of them studied for seven years or more and had family income of two Brazilian basic salaries. The prevalence of HPV DNA was 7.0% (IC95%: 4.9-9.2%), being higher (9.9%) in the group of women aging from 18 to 34 years and lower (4.2%) in the group of women aging between 35 and 59 years. A higher risk of HPV infection was estimated for women informing to have four or more sexual partners in life (adjusted OR = 2.8; IC95%1.12-7.01). Family income and age were inversely related to the risk of developing HPV infection. Conclusion: The HPV prevalence here observed is lower as compared with those observed in other studies accomplished in Latin America. This population presents peculiar characteristics, such as those related to education, income, number of sexual partners and Pap smear history, which may explain the fact of being at lower risk for HPV infection.

17

Ilana Nogueira Bezerra: Eating out of home and obesity: a Brazilian nationwide survey Ilana Nogueira Bezerra, Rosely Sichieri Objective: The aim of this study was to investigate the relationship between out-of-home (OH) eating and the prevalence of overweight and obesity among adults in Brazil. Methods: Data was based on the 2002-2003 Household Budget Survey (48,470 households) conducted by The Brazilian Census Bureau. Foods and drinks purchased for out-of-home eating during a one-week period were recorded by each participant. We considered OH eating the purchase of at least one food or drink item for OH consumption during this period. We classified items as: soft drinks, deep-fried snacks, fast foods, sweets, and sit-down meals. Only data from urban areas were included. Individuals less then 25 and over 65 years old were excluded, yielding a final sample size of 57,524 subjects (26,225 men and 31,299 women). All statistical analyses were performed separately for men and women, using the survey procedure in the statistical software SAS. Frequencies and prevalences were weighted, and statistical analyses took into account the sample design effect. Results: The weighted prevalence of OH eating was 39.6%. Overall, OH eating was positively associated with overweight (OR=1.21; 95% CI 1.10-1.33) and obesity (OR=1.35; 95% CI 1.16-1.57) among men, but not among women. Sit-down meals and soft drinks were the most frequently reported food groups. Both were positively associated with overweight (OR=1.34 for meals; OR= 1.17 for soft drinks, P < 0.05) and obesity (OR=1.51 for meals; OR=1.39 for soft drinks, P < 0.05) among men, but negatively associated with overweight and obesity among women. Conclusion: OH eating was associated with overweight and obesity only among men, whereas, among women, eating sit-down meals OH was protective for obesity, suggesting that women make healthier food choices when they eat out of home.

18

Karine de Lima Sírio Boclin: Racial/ethnic and socioeconomic inequalities in occurrence of uterine myomas Karine de Lima Sírio Boclin, Eduardo Faerstein Introduction: Uterine myomas are the most frequent benign neoplasm of the female reproductive system. Although not related with mortality, these tumors cause significant morbidity for many women. Some US epidemiologic studies show that black women have higher prevalence than other racial-ethnic groups in all ages, in addition to more tumors, more severe symptoms, younger age at diagnosis and higher hysterectomy rates. However, the nature of the association between racial-ethnic and social factors with the occurrence of uterine myomas remains unexplained, and was never studied in Brazil. Objective: To investigate the racial/ethnic and socioeconomic inequalities in the occurrence of uterine myomas among Brazilian women. Methods: The analyses are based on data from 1,820 female civil servants of a university in Rio de Janeiro, Brazil, who participated in two stages of baseline data collection (1999-2001) of a longitudinal study (the Pró-Saúde Study). The outcome variable is the self-reported history of medical diagnosis of uterine myoma. The other variables are: color/race (open-ended) self-identification; markers of participants' early and later-life socioeconomic position; life style variables; atherogenic factors and reproductive life characteristics. Prevalence and crude and adjusted prevalence ratios were estimated from Poisson regression models. Results: According to preliminary analyses 19.4% of the white women, 22.9% of the 'mulato' (or brown) and 32.8% of the blacks reported having had a medical diagnosis of uterine myoma. After adjustment for other co-variables the prevalence ratio of brown women was 1,0 (95% CI: 0,7 – 1,5) and for black women was 1,4 (95 CI %: 1,0 – 2,0) compared with white. According to these preliminary results, being of black color/race is associated with an increased occurrence of uterine myoma in this population. Other known risk factors do not explain the observed association.

19

João Luiz Dornelles Bastos: Racial inequalities in health: measuring self-reported experiences of discrimination in Brazil João Luiz Dornelles Bastos (PhD student in epidemiology), Aluísio Jardim Dornellas de Barros (supervisor), Eduardo Faerstein (co-supervisor) Since the middle of 20th century, social inequalities in health are increasingly becoming part of the mainstream public health research agenda. Besides its theoretical and methodological controversies, the investigation of racial inequalities in health stands as a promising line of research in the area. As a means of contributing to a fuller understanding of what determines racial inequalities in health, some scholars have developed methodological strategies to measure self-reported racial discrimination. The objective of this project is to develop a scale to assess self-reported experiences of discrimination in adults from Brazil. The scale's items will be selected/generated through a literature review of existing instruments as well as focus groups discussions. Next, experts in the study of racial relations in Brazil will assess these items. A preliminary version of the scale will be pilot-tested for item refinement. Finally, the scale will be applied among undergraduate students from the State University of Rio de Janeiro, so that validity and reliability analyses can be carried out. The development of such a discrimination scale may provide a fuller accounting of what drives population patterns of health in Brazil.

20

Mikael Rostila: Birds of a feather flock together – and fall ill? Migrant homophily and health in Sweden Although studies suggest ethnic health inequalities, few have expressed interest in the origins of such disparities. Homophily refers to the fact that people form relationships principally with those who are similar to themselves. As homophily influences people’s attitudes, information and interactions, it might be an important property of migrants’ social networks, with implications on their health. The aim of this study is to examine the health consequences arising from participation in homogenous migrant networks in Sweden. Initial analyses suggest higher mortality rates among several groups of migrants. The findings further suggest that migrant homophily is a prominent feature of social networks and that migrants in homogenous networks have poorer health than do migrants with natives included in their networks. However, analyses suggest that it is merely that migrants with homogenous and closed networks report poorer health, and that unhealthy behaviours and disadvantaged socioeconomic conditions may account for much of their excess risk. It is suggested that network closure may reinforce and maintain unhealthy norms and socioeconomic conditions in homogenous migrant networks. Overall, the findings suggest that negative health externalities may arise from networks and that it may be important to consider network properties when explaining and preventing ethnic health disparities.

21

List of participants 1 Alba Maria Zaluar amz84@globo.com 2 Aldair Jose de Oliveira aldair_alir@ig.com.br 3 Alessandra Oliveira alessadias@ims.uerj.br 4 Alexandre Dos Santos Brito britoalexandre@ims.uerj.br 5 Alice Helena Nóra Pacheco Gagliano alicehelenapacheco@yahoo.com.br 6 Amanda de Moura Souza amanadamoura@ims.uerj.br 7 Ana Carolina Reiff Vieira carolreiff@gmail.com 8 Ana Paula Pires dos Santos ana.paulapires@uol.com.br 9 Anna Maria Sales anna@radnet.com.br

10 Antonio Ponce de Leon ponce@ims.uerj.br 11 Carlos Castro Luz 12 Carlos Eduardo Raymundo caducer@gmail.com 13 Claudia Leite Moraes clmoraes@ims.uerj.br 14 Cláudia Soares claudia.sr@globo.com 15 Cláudia Souza de Lopes lopes@ims.uerj.br 16 Eduardo Faerstein eduardof@uerj.br 17 Fabio da Silva Gomes fabiodasilvagomes@gmail.com 18 Fernanda Nogueira fernandanogueira5@hotmail.com 19 Flávia Dos Santos Barbosa barbosaflavia@ims.uerj.br 20 Flavio Gomes Ferreira Pinto flaviogfp@hotmail.com 21 Gulnar Azevedo e Silva gulnar@ims.uerj.br 22 Gustavo Lobato lobato@iff.fiocruz.br 23 Ilana Nogueira Bezerra ilana.bezerra@yahoo.com.br 24 Jacqueline Furtado Vital jacqueline.furtado@ims.uerj.br 25 Jacyara Brito Macedo jacyara@uerj.br 26 Janaina Carvalho jmac.med@gmail.com 27 João Luiz Dornelles Bastos joao@pilotis.com.br 28 Johan Fritzell johan.fritzell@chess.su.se 29 Karine de Lima Sírio Boclin kboclin@hotmail.com 30 Kenneth Rochel de Camargo kenneth@uerj.br 31 Luciana Tricai Cavalini lutricav@vm.uff.br 32 Marcia Guimarães de Mello Alves marciagmalves@gmail.com 33 Maria do Carmo Esteves da Costa mcarmo@inca.gov.br 34 Maria Kölegård maria.k.stjarne@chess.su.se 35 Mauricio Berger mauricio.berger@gmail.com 36 Michael Reichenheim michael@ims.uerj.br 37 Mikael Rostila mikael.rostial@chess.su.se 38 Ney Armando de Mello Meziat Filho neymeziat@gmail.com 39 Odaleia B Aguiar odaleiab@hotmail.com 40 Osmar da Silva Barros Junior majosmarjr@yahoo.com.br 41 Paula Florence paulaflorence@gmail.com

22

42 Rita Adriana Gomes de Souza ritadriana@ims.uerj.br 43 Roger Keller rk_celeste@socarb.su.se 44 Rosanne Harter Griep rohgriep@ioc.fiocruz.br 45 Roseli Sichieri sichieri@ims.uerj.br 46 Susanna Toivanen susanna.toivanen@chess.su.se 47 Vanessa dos Reis de Souza vrsouza2003@yahoo.com.br 48 Washington Leite Junger wjunger@gmail.com 49 Yara H. M. Hökerberg yarahahr@ipec.fiocruz.br