Measuring socioeconomic position (SEP) in population chronic disease and risk factor surveillance systems is essential for monitoring changes in socioeconomic inequities in health over time. A life course approach in epidemiology considers the long-term effects of physical and social exposures during gestation, childhood, adolescence, and later adult life on health. Previous studies provide evidence that socioeconomic factors at different stages of the life course influence current health status. Measures of SEP during early life to supplement existing indicators of current SEP are required to more adequately explain the contribution of socioeconomic factors to health status and monitor health inequities. The aim of this thesis was to examine how a life course perspective could enhance the monitoring of SEP in chronic disease and risk factor surveillance systems. The thesis reviewed indicators of early life SEP used in previous research, determined indicators of early life SEP that may be useful in South Australian surveillance systems, and examined the association of SEP over the life course and self-rated health in adulthood across different population groups to demonstrate that inclusion of indicators of early life SEP in surveillance systems could allow health inequities to be monitored among socially mobile and stable groups. A variety of indicators, such as parents’ education level and occupation, and financial circumstances and living conditions during childhood, have been used in different study designs in many countries. Indicators of early life SEP used to monitor trends in the health and SEP of populations over time, and to analyse long-term effects of policies on the changing health of populations, need to be feasible to measure retrospectively, and relevant to the historical, geographical and sociocultural context in which the surveillance system is operating. Retrospective recall of various indicators of early life SEP was examined in a telephone survey of a representative South Australian sample of adults. The highest proportions of missing data were observed for maternal grandfather’s occupation, and mother’s and father’s highest education level. Family structure, housing tenure, and family financial situation when the respondent was aged ten, and mother and father’s main occupation had lower item non-response. Respondents with missing data on early life SEP indicators were disadvantaged in terms of current SEP compared to those who provided this information. The differential response to early life SEP questions according to current circumstances has implications for chronic disease surveillance examining the life course impact of socioeconomic disadvantage. While face-to-face surveys are considered the gold standard of interviewing techniques, computer-assisted telephone interviewing is often preferred for cost and convenience. Recall of father’s and mother’s highest education level in the telephone survey was compared to that obtained in a face-to-face interview survey. The proportion of respondents who provided information about their father’s and mother’s highest education level was significantly higher in the face-to-face interview than in the telephone interview. Survey mode, however, did not influence the finding that respondents with missing data for parents’ education were more likely to be socioeconomically disadvantaged. Alternative indicators of early life SEP, such as material and financial circumstances, are likely to be more appropriate than parents’ education for life course analyses of health inequities using surveillance data. Questions about family financial situation and housing tenure during childhood and adulthood asked in the cross-sectional telephone survey were used to examine the association of SEP over the life course with self-rated health in adulthood. Disadvantaged SEP during both childhood and adulthood and upward social mobility in financial situation were associated with a reduced prevalence of excellent or very good health, although this relationship varied across gender, rurality, and country of birth groups. Trend data from a chronic disease and risk factor surveillance system indicated that socioeconomic disadvantage in adulthood was associated with poorer self-rated health. The surveillance system, however, does not currently contain any measures of early life SEP. Overlaying the social mobility variables on the surveillance data indicated how inequities in health could be differentiated in greater detail if early life SEP was measured in addition to current SEP. Inclusion of life course SEP measures in surveillance will enable monitoring of health inequities trends among socially mobile and stable groups. Life course measures are an innovative way to supplement other SEP indicators in surveillance systems. Considerable information can be gained with the addition of a few questions. This will provide further insight into the determinants of health and illness and enable improved monitoring of the effects of policies and interventions on health inequities and intergenerational disadvantage. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1367190 / Thesis (Ph.D.) - University of Adelaide, School of Population Health and Clinical Practice, 2009

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Esta tese é composta por três artigos que permitiram avaliar o efeito da exposição ao tabagismo durante a gestação e no início da infância sobre o crescimento linear e ganho de peso do nascimento à adolescência, além de verificar o efeito do nível socioeconômico no início da infância e da mobilidade social sobre a adiposidade até a adolescência. Foram utilizados para este fim os dados de uma coorte de crianças nascidas entre 1994 e 1999 na cidade de Cuiabá-MT. Essas crianças fizeram parte de um estudo de base populacional realizado na cidade de Cuiabá, entre 1999 e 2000, com 2405 crianças (0 a 5 anos) e foram selecionadas aleatoriamente em unidades básicas de saúde quando da vacinação. As mães foram entrevistadas após a vacinação, quando foram obtidos dados relativos à exposição ao tabagismo gestacional, tabagismo passivo, nível socioeconômico das famílias e dados antropométricos. Entre 2009 e 2011, após aproximadamente 11 anos, essas crianças foram localizadas por meio do Censo Escolar e então 1716 adolescentes entre 10 e 17 anos de idade (71,4% da população) foram reavaliados nas escolas da rede pública e privada de Cuiabá, de 18 municípios do estado e outras 5 capitais do país. A análise por modelos lineares de efeitos mistos permitiu verificar a mudança de estatura e Índice de Massa Corporal (IMC) entre o nascimento e a adolescência. O primeiro e o segundo artigo desta tese avaliaram o efeito da exposição ao tabagismo materno durante a gestação e no início da infância sobre o crescimento linear e o IMC entre o nascimento e a adolescência. Crianças expostas ao tabagismo materno durante a gestação e no início da infância apresentaram menor estatura desde o nascimento até a adolescência quando comparadas às crianças não expostas. Quanto à adiposidade, entre o nascimento e a infância a mudança do IMC foi similar entre as crianças expostas e não expostas ao tabagismo materno, porém, entre a infância e a adolescência, aquelas expostas apenas durante a gestação mostraram maior ganho de IMC. Em conjunto, os dados corroboram o efeito deletério do tabagismo sobre o crescimento, efeito já bastante estudado, mas também indicam que avaliar e comparar exposição gestacional com pós-gestacional é importante, dado que seus efeitos parecem ser diferentes. O terceiro artigo avaliou o efeito do nível socioeconômico no início da infância e da mobilidade social entre a infância e a adolescência sobre o IMC do nascimento à adolescência. Para avaliar o nível socioeconômico, as famílias foram classificadas em nível econômico alto, médio e baixo, a partir do Critério de Classificação Econômica Brasil. Foi observada expressiva mobilidade social na população, principalmente entre os de menor nível econômico. Houve maior aumento do IMC entre o nascimento e a adolescência entre aqueles de maior nível econômico na infância e aqueles que permaneceram nas classes mais elevadas, indicando que a posição inicial foi o maior determinante das mudanças observadas no IMC. / This thesis consists of three papers that allowed us to evaluate the effect of exposure to smoking during pregnancy and early childhood on linear growth and weight gain from birth to adolescence, and to verify the effect of socioeconomic position in early childhood and social mobility on adiposity until adolescence. For this purpose, we used the data of a cohort of children born between 1994 and 1999 in the city of Cuiabá. They were part of a population-based study conducted with 2,405 children (05 years) who attended one of the ten basic health units selected for vaccination in the city of Cuiabá between 1999 and 2000. Between 2009 and 2011, after approximately 11 years, these children were located through the School Census; thus, 1,716 adolescents aged 10 to 17 years old (71.4% of the population) were reassessed in the public and private schools of Cuiabá, of 18 counties in the state and five other capitals in the country. The mothers were interviewed after vaccination to obtain data regarding their exposure to smoking during pregnancy, passive smoking, socioeconomic status, and anthropometric information. The linear mixed-effects analysis showed the change in height and body mass index (BMI) from birth to adolescence. The first and second articles in this thesis evaluate the effect of exposure to maternal smoking during pregnancy and early childhood on linear growth and BMI from birth to adolescence. Children exposed to maternal smoking during pregnancy and early childhood were shorter from birth to adolescence than the children not exposed. As for adiposity between birth and childhood, the BMI change was similar between children exposed and not exposed to maternal smoking; however, between childhood and adolescence, those exposed only during gestation showed greater gains in BMI. Together, these data not only confirm the deleterious effect of smoking on growth, which has been extensively studied, but also indicate that it is important to evaluate and compare gestational to post-pregnancy exposure because their effects seem to be different. The third paper examines the effect of socioeconomic status in early childhood and social mobility from childhood to adolescence on BMI from birth to adolescence. To assess socioeconomic position, the families were classified into high, medium, and low economic levels based on the Brazil Criterion of economic classification. We observed significant social mobility in the population, especially among the lower economic group. We observed a greater increase in BMI from birth to adolescence among those from higher socioeconomic position in childhood and those who remained in the higher classes, indicating that the starting position was the major determinant of the observed changes in BMI.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / Esta dissertação buscou uma apreciação crítica e pragmática da relação entre desigualdade socioeconômica e obesidade abdominal, em resposta a proposição internacional de monitoramento das desigualdades em saúde e a escassez de estudos desta natureza relativos à obesidade abdominal. Dois artigos foram elaborados a fim de estimar o grau de desigualdade educacional na ocorrência de obesidade abdominal e revisar os estudos de associação entre posição socioeconômica e obesidade abdominal. O primeiro artigo utilizou o índice angular de desigualdade e o índice relativo de desigualdade em dados seccionais de 3.117 participantes da linha de base do Estudo Pró-Saúde, 1999-2001, e o segundo artigo abarcou os resultados de estudos conduzidos em população adulta no Brasil. Os índices de desigualdade resumiram a tendência monotônica e inversa observada entre escolaridade e obesidade abdominal na população feminina, proporcionando estimativas quantitativas desta desigualdade (artigo 1). Em concordância, observou-se que a associação entre indicadores de posição socioeconômica e obesidade abdominal foi majoritariamente inversa entre as mulheres, principalmente com relação à escolaridade, e estatisticamente não significativa entre os homens (artigo 2). Tal cenário epidemiológico evidencia que a obesidade abdominal tem afetado desproporcionalmente as mulheres de posição socioeconômica mais baixa e que a desigualdade de gênero na prevalência de obesidade abdominal tende a aumentar com menor posição socioeconômica. Em suma, a presente dissertação visou à produção de conhecimento epidemiológico relevante ao enfrentamento das desigualdades em saúde, com o objetivo premente de subsidiar políticas públicas de fato realizáveis e individualmente aceitáveis. / This dissertation sought to a critical and pragmatic assessment of the relationship between socioeconomic inequality and abdominal obesity in response to international proposition of health inequalities monitoring and to lack of studies of this nature relating to abdominal obesity. Two articles were prepared to -estimate the level of educational inequality in the occurrence of abdominal obesity, and review the association studies between socioeconomic position and abdominal obesity. The first article used the slope index of inequality and the relative index of inequality in the sectional data of 3.117 participants in the baseline of the Pró-Saúde Study, 1999-2001; and the second article encompassed the results of the studies conducted in the adult population in Brazil. The inequality indexes summarized strictly monotonic and inverse trend between educational achievement and abdominal obesity in the female population, providing quantitative estimates of this inequality (Article 1). Accordingly, we found that the association between socioeconomic position indicators and abdominal obesity was mostly reversed among women, especially regarding education, and statistically not significant among men (Article 2). This epidemiological scenario shows that abdominal obesity has disproportionately affected women of the lower socioeconomic position, and gender inequality in the prevalence of abdominal obesity tends to increase with lower socioeconomic position. In short, this dissertation aimed at the production of relevant epidemiological knowledge to addressing health inequalities, with the targeted of subsidize public policies feasible and individually acceptable.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Os miomas uterinos (MU) são considerados os tumores mais comuns do sistema reprodutor feminino. Estudos norte-americanos demonstram que mulheres negras são mais acometidas pelos MU que as de outros grupos étnico-raciais. No entanto, as causas da desigualdade racial na ocorrência dos tumores permanecem desconhecidas e possíveis mecanismos são pouco explorados na literatura. Em outra direção, devido às características dos MU (crescimento lento e longo período de latência) parte considerável dos estudos epidemiológicos utilizam um delineamento transversal, o que pode gerar problemas metodológicos, como os relacionados à utilização da idade coletada transversalmente (posteriormente a ocorrência dos MU) como proxy da idade do surgimento dos tumores. Assim, este trabalho de tese foi dividido em três partes, como se segue. A primeira, com características descritivas, teve por objetivo estimar a ocorrência de MU autorelatados segundo categorias demográficas e sócio-econômicas na população de estudo (compôs o artigo 1). A segunda, com componente analítico, propôs-se a avaliar o papel da PSE ao longo da vida como mediadora do efeito da cor/raça na ocorrência de MU auto-relatados (compôs o artigo 2). A terceira, com caráter metodológico, teve por objetivo comparar medidas de associação, entre variáveis aferidas transversalmente, em análises que incluem a co-variável idade no momento da coleta de dados e análises que consideram a idade ao diagnóstico dos MU (compôs o artigo 3). Para tanto, foram analisados dados transversais da população feminina participante das duas etapas da linha de base do Estudo Pró-Saúde, referentes à história auto-relatada de diagnóstico médico de MU e ainda a características sócio-demográficas, da vida reprodutiva e de acesso a serviços de saúde. Os resultados evidenciaram o aumento de ocorrência de MU em mulheres de maior idade e com a cor da pele mais escura (artigo 1); que a PSE ao longo da vida não medeia as associações entre cor/raça e MU (artigo 2); e que apesar das diferenças de pequena magnitude, a idade referida no momento da coleta de dados parece ser menos indicada para fins de especificação dos modelos analíticos do que a idade ao diagnóstico dos MU(artigo 3). / The uterine myomas (UM) are considered the most frequent benign neoplasm of the female reproductive system. U.S. studies showed that UM occur more frequently among black women, but the nature of this association remains largely unexplained in the literature. In another direction, due to the characteristics of the UM (slow growth and latency period) considerable amount of epidemiological studies use a cross-sectional design, which can lead to methodological problems such as those related to the use of age collected transversally (later the occurrence of UM) as a proxy for age of onset of tumors. This thesis was divided into three parts, as follows. The first, descriptive, aimed to estimate the occurrence of self-reported UM by demographic and socio-economic characteristics in the study population (article 1). The second, whit an analytical component, aimed to evaluate the role of life-course SEP like mediator of the effect of color/race in the occurrence of self-reported UM (article 2). The third, with a methodological nature, aimed comparing measures of association between variables collected transversally, in analysis that include the covariate age at the time of data collection and analysis that considered the age at diagnosis of UM (article 3).For this, we analyzed cross-sectional data from selfadministered questionnaires completed by female civil servants at a Rio de Janeiro university during the baseline data collection of the Pró-Saúde Study. The analyzed variables were: self-reported history of medical diagnosis of UM, UM with symptoms prior to diagnosis, hysterectomy due to UM and also the socio-demographic characteristics, reproductive life and health care access variables. The results showed that UM risk increased with the age and darkening of skin color, and the lifecourse SEP did not mediate this association (articles 1 and 2); and that despite differences of small magnitude, it seems that the age at the time of data collection is less recommended than the age at diagnosis in cross-sectional analysis (article 3).

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / Esta dissertação buscou uma apreciação crítica e pragmática da relação entre desigualdade socioeconômica e obesidade abdominal, em resposta a proposição internacional de monitoramento das desigualdades em saúde e a escassez de estudos desta natureza relativos à obesidade abdominal. Dois artigos foram elaborados a fim de estimar o grau de desigualdade educacional na ocorrência de obesidade abdominal e revisar os estudos de associação entre posição socioeconômica e obesidade abdominal. O primeiro artigo utilizou o índice angular de desigualdade e o índice relativo de desigualdade em dados seccionais de 3.117 participantes da linha de base do Estudo Pró-Saúde, 1999-2001, e o segundo artigo abarcou os resultados de estudos conduzidos em população adulta no Brasil. Os índices de desigualdade resumiram a tendência monotônica e inversa observada entre escolaridade e obesidade abdominal na população feminina, proporcionando estimativas quantitativas desta desigualdade (artigo 1). Em concordância, observou-se que a associação entre indicadores de posição socioeconômica e obesidade abdominal foi majoritariamente inversa entre as mulheres, principalmente com relação à escolaridade, e estatisticamente não significativa entre os homens (artigo 2). Tal cenário epidemiológico evidencia que a obesidade abdominal tem afetado desproporcionalmente as mulheres de posição socioeconômica mais baixa e que a desigualdade de gênero na prevalência de obesidade abdominal tende a aumentar com menor posição socioeconômica. Em suma, a presente dissertação visou à produção de conhecimento epidemiológico relevante ao enfrentamento das desigualdades em saúde, com o objetivo premente de subsidiar políticas públicas de fato realizáveis e individualmente aceitáveis. / This dissertation sought to a critical and pragmatic assessment of the relationship between socioeconomic inequality and abdominal obesity in response to international proposition of health inequalities monitoring and to lack of studies of this nature relating to abdominal obesity. Two articles were prepared to -estimate the level of educational inequality in the occurrence of abdominal obesity, and review the association studies between socioeconomic position and abdominal obesity. The first article used the slope index of inequality and the relative index of inequality in the sectional data of 3.117 participants in the baseline of the Pró-Saúde Study, 1999-2001; and the second article encompassed the results of the studies conducted in the adult population in Brazil. The inequality indexes summarized strictly monotonic and inverse trend between educational achievement and abdominal obesity in the female population, providing quantitative estimates of this inequality (Article 1). Accordingly, we found that the association between socioeconomic position indicators and abdominal obesity was mostly reversed among women, especially regarding education, and statistically not significant among men (Article 2). This epidemiological scenario shows that abdominal obesity has disproportionately affected women of the lower socioeconomic position, and gender inequality in the prevalence of abdominal obesity tends to increase with lower socioeconomic position. In short, this dissertation aimed at the production of relevant epidemiological knowledge to addressing health inequalities, with the targeted of subsidize public policies feasible and individually acceptable.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Os miomas uterinos (MU) são considerados os tumores mais comuns do sistema reprodutor feminino. Estudos norte-americanos demonstram que mulheres negras são mais acometidas pelos MU que as de outros grupos étnico-raciais. No entanto, as causas da desigualdade racial na ocorrência dos tumores permanecem desconhecidas e possíveis mecanismos são pouco explorados na literatura. Em outra direção, devido às características dos MU (crescimento lento e longo período de latência) parte considerável dos estudos epidemiológicos utilizam um delineamento transversal, o que pode gerar problemas metodológicos, como os relacionados à utilização da idade coletada transversalmente (posteriormente a ocorrência dos MU) como proxy da idade do surgimento dos tumores. Assim, este trabalho de tese foi dividido em três partes, como se segue. A primeira, com características descritivas, teve por objetivo estimar a ocorrência de MU autorelatados segundo categorias demográficas e sócio-econômicas na população de estudo (compôs o artigo 1). A segunda, com componente analítico, propôs-se a avaliar o papel da PSE ao longo da vida como mediadora do efeito da cor/raça na ocorrência de MU auto-relatados (compôs o artigo 2). A terceira, com caráter metodológico, teve por objetivo comparar medidas de associação, entre variáveis aferidas transversalmente, em análises que incluem a co-variável idade no momento da coleta de dados e análises que consideram a idade ao diagnóstico dos MU (compôs o artigo 3). Para tanto, foram analisados dados transversais da população feminina participante das duas etapas da linha de base do Estudo Pró-Saúde, referentes à história auto-relatada de diagnóstico médico de MU e ainda a características sócio-demográficas, da vida reprodutiva e de acesso a serviços de saúde. Os resultados evidenciaram o aumento de ocorrência de MU em mulheres de maior idade e com a cor da pele mais escura (artigo 1); que a PSE ao longo da vida não medeia as associações entre cor/raça e MU (artigo 2); e que apesar das diferenças de pequena magnitude, a idade referida no momento da coleta de dados parece ser menos indicada para fins de especificação dos modelos analíticos do que a idade ao diagnóstico dos MU(artigo 3). / The uterine myomas (UM) are considered the most frequent benign neoplasm of the female reproductive system. U.S. studies showed that UM occur more frequently among black women, but the nature of this association remains largely unexplained in the literature. In another direction, due to the characteristics of the UM (slow growth and latency period) considerable amount of epidemiological studies use a cross-sectional design, which can lead to methodological problems such as those related to the use of age collected transversally (later the occurrence of UM) as a proxy for age of onset of tumors. This thesis was divided into three parts, as follows. The first, descriptive, aimed to estimate the occurrence of self-reported UM by demographic and socio-economic characteristics in the study population (article 1). The second, whit an analytical component, aimed to evaluate the role of life-course SEP like mediator of the effect of color/race in the occurrence of self-reported UM (article 2). The third, with a methodological nature, aimed comparing measures of association between variables collected transversally, in analysis that include the covariate age at the time of data collection and analysis that considered the age at diagnosis of UM (article 3).For this, we analyzed cross-sectional data from selfadministered questionnaires completed by female civil servants at a Rio de Janeiro university during the baseline data collection of the Pró-Saúde Study. The analyzed variables were: self-reported history of medical diagnosis of UM, UM with symptoms prior to diagnosis, hysterectomy due to UM and also the socio-demographic characteristics, reproductive life and health care access variables. The results showed that UM risk increased with the age and darkening of skin color, and the lifecourse SEP did not mediate this association (articles 1 and 2); and that despite differences of small magnitude, it seems that the age at the time of data collection is less recommended than the age at diagnosis in cross-sectional analysis (article 3).

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Esta tese é composta por três artigos que permitiram avaliar o efeito da exposição ao tabagismo durante a gestação e no início da infância sobre o crescimento linear e ganho de peso do nascimento à adolescência, além de verificar o efeito do nível socioeconômico no início da infância e da mobilidade social sobre a adiposidade até a adolescência. Foram utilizados para este fim os dados de uma coorte de crianças nascidas entre 1994 e 1999 na cidade de Cuiabá-MT. Essas crianças fizeram parte de um estudo de base populacional realizado na cidade de Cuiabá, entre 1999 e 2000, com 2405 crianças (0 a 5 anos) e foram selecionadas aleatoriamente em unidades básicas de saúde quando da vacinação. As mães foram entrevistadas após a vacinação, quando foram obtidos dados relativos à exposição ao tabagismo gestacional, tabagismo passivo, nível socioeconômico das famílias e dados antropométricos. Entre 2009 e 2011, após aproximadamente 11 anos, essas crianças foram localizadas por meio do Censo Escolar e então 1716 adolescentes entre 10 e 17 anos de idade (71,4% da população) foram reavaliados nas escolas da rede pública e privada de Cuiabá, de 18 municípios do estado e outras 5 capitais do país. A análise por modelos lineares de efeitos mistos permitiu verificar a mudança de estatura e Índice de Massa Corporal (IMC) entre o nascimento e a adolescência. O primeiro e o segundo artigo desta tese avaliaram o efeito da exposição ao tabagismo materno durante a gestação e no início da infância sobre o crescimento linear e o IMC entre o nascimento e a adolescência. Crianças expostas ao tabagismo materno durante a gestação e no início da infância apresentaram menor estatura desde o nascimento até a adolescência quando comparadas às crianças não expostas. Quanto à adiposidade, entre o nascimento e a infância a mudança do IMC foi similar entre as crianças expostas e não expostas ao tabagismo materno, porém, entre a infância e a adolescência, aquelas expostas apenas durante a gestação mostraram maior ganho de IMC. Em conjunto, os dados corroboram o efeito deletério do tabagismo sobre o crescimento, efeito já bastante estudado, mas também indicam que avaliar e comparar exposição gestacional com pós-gestacional é importante, dado que seus efeitos parecem ser diferentes. O terceiro artigo avaliou o efeito do nível socioeconômico no início da infância e da mobilidade social entre a infância e a adolescência sobre o IMC do nascimento à adolescência. Para avaliar o nível socioeconômico, as famílias foram classificadas em nível econômico alto, médio e baixo, a partir do Critério de Classificação Econômica Brasil. Foi observada expressiva mobilidade social na população, principalmente entre os de menor nível econômico. Houve maior aumento do IMC entre o nascimento e a adolescência entre aqueles de maior nível econômico na infância e aqueles que permaneceram nas classes mais elevadas, indicando que a posição inicial foi o maior determinante das mudanças observadas no IMC. / This thesis consists of three papers that allowed us to evaluate the effect of exposure to smoking during pregnancy and early childhood on linear growth and weight gain from birth to adolescence, and to verify the effect of socioeconomic position in early childhood and social mobility on adiposity until adolescence. For this purpose, we used the data of a cohort of children born between 1994 and 1999 in the city of Cuiabá. They were part of a population-based study conducted with 2,405 children (05 years) who attended one of the ten basic health units selected for vaccination in the city of Cuiabá between 1999 and 2000. Between 2009 and 2011, after approximately 11 years, these children were located through the School Census; thus, 1,716 adolescents aged 10 to 17 years old (71.4% of the population) were reassessed in the public and private schools of Cuiabá, of 18 counties in the state and five other capitals in the country. The mothers were interviewed after vaccination to obtain data regarding their exposure to smoking during pregnancy, passive smoking, socioeconomic status, and anthropometric information. The linear mixed-effects analysis showed the change in height and body mass index (BMI) from birth to adolescence. The first and second articles in this thesis evaluate the effect of exposure to maternal smoking during pregnancy and early childhood on linear growth and BMI from birth to adolescence. Children exposed to maternal smoking during pregnancy and early childhood were shorter from birth to adolescence than the children not exposed. As for adiposity between birth and childhood, the BMI change was similar between children exposed and not exposed to maternal smoking; however, between childhood and adolescence, those exposed only during gestation showed greater gains in BMI. Together, these data not only confirm the deleterious effect of smoking on growth, which has been extensively studied, but also indicate that it is important to evaluate and compare gestational to post-pregnancy exposure because their effects seem to be different. The third paper examines the effect of socioeconomic status in early childhood and social mobility from childhood to adolescence on BMI from birth to adolescence. To assess socioeconomic position, the families were classified into high, medium, and low economic levels based on the Brazil Criterion of economic classification. We observed significant social mobility in the population, especially among the lower economic group. We observed a greater increase in BMI from birth to adolescence among those from higher socioeconomic position in childhood and those who remained in the higher classes, indicating that the starting position was the major determinant of the observed changes in BMI.

The aim for this thesis was to identify prognostic factors for chronic diabetes complications that exist at the time of diabetes diagnosis. Low level of education (<12 years) and low income (<60% of median) was found to increase the risk to have high (>70 mmol/mol) HbA1c at the time of diagnosis with 34 % and 35 %, respectively. Prevalence of diabetic retinopathy (DR) was 12% in a cohort of patients newly diagnosed with diabetes. Diabetic macular edema was present in 11% of patients with type 2 diabetes (T2D) and 13% of those with Latent Autoimmune Diabetes in Adults (LADA). Low beta cell function and low level of education increased the risk for DR with 110% and 43%, respectively. For every unit of increase in body mass index, the risk for DR was reduced by 3%. The cellular immunology of LADA patients was a mixture of that observed in both type 1 (T1D) and T2D patients. Compared to patients with T1D, LADA patients had more B-regulatory lymphocytes and antigen presenting cells capable of producing interleukine-35. This indicates a higher anti-inflammatory capacity in LADA patients compared to type T1D patients. By imputing age, body mass index, HbA1c at diagnosis, beta cell function and insulin resistance in a cluster analysis, five distinct diabetes clusters were identified. The four clusters representing T2D patients differed in incidence of DR, nephropathy and non-alcoholic fatty liver disease. This was replicated with similar results in three geographically separate populations. By studying socioeconomic background and factors present at the time of diagnosis we can better predict prognosis for chronic diabetes complications. These findings may facilitate better-targeted diabetes screening programs and more individually tailored treatment regimes.

L'activité physique améliore la santé, mais seulement 4.8% des Canadiens atteignent le niveau recommandé. La position socio-économique est un des déterminants de l'activité physique les plus importants. Elle est associée à l’activité physique de manière transversale à l’adolescence et à l’âge adulte. Cette thèse a tenté de déterminer s'il y a une association à long terme entre la position socio-économique au début du parcours de vie et l’activité physique à l’âge adulte. S'il y en avait une, un deuxième objectif était de déterminer quel modèle théorique en épidémiologie des parcours de vie décrivait le mieux sa forme. Cette thèse comprend trois articles: une recension systématique et deux recherches originales. Dans la recension systématique, des recherches ont été faites dans Medline et EMBASE pour trouver les études ayant mesuré la position socio-économique avant l'âge de 18 ans et l'activité physique à ≥18 ans. Dans les deux recherches originales, la modélisation par équations structurelles a été utilisée pour comparer trois modèles alternatifs en épidémiologie des parcours de vie: le modèle d’accumulation de risque avec effets additifs, le modèle d’accumulation de risque avec effet déclenché et le modèle de période critique. Ces modèles ont été comparés dans deux cohortes prospectives représentatives à l'échelle nationale: la 1970 British birth cohort (n=16,571; première recherche) et l’Enquête longitudinale nationale sur les enfants et les jeunes (n=16,903; deuxième recherche). Dans la recension systématique, 10 619 articles ont été passés en revue par deux chercheurs indépendants et 42 ont été retenus. Pour le résultat «activité physique» (tous types et mesures confondus), une association significative avec la position socio-économique durant l’enfance fut trouvée dans 26/42 études (61,9%). Quand seulement l’activité physique durant les loisirs a été considérée, une association significative fut trouvée dans 21/31 études (67,7%). Dans un sous-échantillon de 21 études ayant une méthodologie plus forte, les proportions d’études ayant trouvé une association furent plus hautes : 15/21 (71,4%) pour tous les types et toutes les mesures d’activité physique et 12/15 (80%) pour l’activité physique de loisir seulement. Dans notre première recherche originale sur les données de la British birth cohort, pour la classe sociale, nous avons trouvé que le modèle d’accumulation de risque avec effets additifs s’est ajusté le mieux chez les hommes et les femmes pour l’activité physique de loisir, au travail et durant les transports. Dans notre deuxième recherche originale sur les données canadiennes sur l'activité physique de loisir, nous avons trouvé que chez les hommes, le modèle de période critique s’est ajusté le mieux aux données pour le niveau d’éducation et le revenu, alors que chez les femmes, le modèle d’accumulation de risque avec effets additifs s’est ajusté le mieux pour le revenu, tandis que le niveau d’éducation ne s’est ajusté à aucun des modèles testés. En conclusion, notre recension systématique indique que la position socio-économique au début du parcours de vie est associée à la pratique d'activité physique à l'âge adulte. Les résultats de nos deux recherches originales suggèrent un patron d’associations le mieux représenté par le modèle d’accumulation de risque avec effets additifs. / Physical activity enhances health, yet only 4.8% of adults in Canada meet the recommended level. Among the determinants of physical activity, socioeconomic position shows some of the strongest and most consistent associations. Its association is found cross-sectionally during adolescence and cross-sectionally during adulthood. This thesis aimed to determine if there is a long-lasting, life course association between socioeconomic position early in life and physical activity during adulthood. If there was one, a second goal was to determine which theoretical model in life course epidemiology best described its pattern. This thesis comprises three papers: a systematic review and two original contributions. In the systematic review, Medline and EMBASE were searched for studies that assessed socioeconomic position before age 18 years and physical activity at age ≥18 years. In the two original research papers, structural equation modeling was used to compare three competing models in life course epidemiology: the accumulation of risk model with additive effects, the accumulation of risk model with trigger effect, and the critical period model. Each of the original research paper used data from a large, nationally representative prospective cohort. In the first paper, models were compared in the 1970 British birth cohort (n=16,571). Social class was used as an indicator of socioeconomic position, and physical activity was assessed during leisure time, during transports, and at work. In the second paper, models were compared in the Canadian National Longitudinal Survey of Children and Youth (n=16,903). Education and income were used as indicators of socioeconomic position and physical activity was assessed during leisure-time only. In the systematic review, 10,619 publications were reviewed by two independent investigators and 42 were retained. For outcome "physical activity" (all types and measures), a significant association with socioeconomic position before age 18 years was found in 26/42 studies (61.9%). When the only outcome considered was leisure-time physical activity, a significant association was found in 21/31 studies (67.7%). In a subset of 21 studies with more rigorous methodology, proportions of studies finding an association were higher: 15/21 (71.4%) for all types and measures of physical activity and 12/15 (80%) for leisure-time physical activity only. In our first original research paper, using British data on social class, we found that the accumulation of risk model with additive effects fit the data best in both men and women for all three domains of physical activity studied (leisure time, transports, and work). In our second original research paper, using Canadian data on leisure-time physical activity, we found that in men, the critical period model fit the data best for education and income, while in women, the accumulation of risk model with additive effects fit the data best for income, whereas education did not fit any model. To conclude, our systematic review suggests that socioeconomic position early in life is associated with physical activity during adulthood. Results from our two original research papers indicate that the pattern of its association may be best represented by the accumulation of risk model with additive effects.