Introdução: O excesso de peso (EP, inclui sobrepeso e obesidade) configura um grande problema de saúde pública no país. A compreensão do estado nutricional ao longo do tempo e de fatores associados, como a dieta, além de suas possíveis consequências, como fatores de risco cardiometabólico, é essencial para o desenvolvimento de ações em saúde. Objetivo: Avaliar a prevalência de sobrepeso e obesidade na população da cidade de São Paulo, assim como modificações na dieta da população entre os anos de 2003 e 2015 e sua associação com fatores de risco cardiometabólico. Métodos: Foram utilizados dados do estudo transversal de base populacional ISA - Capital dos anos de 2003, 2008 e 2015, referentes à amostra probabilística de residentes do município de São Paulo com 12 anos ou mais, de ambos os sexos, totalizando 8.682 indivíduos. Dados socioeconômicos, antropométricos, de estilo de vida e recordatório alimentar de 24 horas foram coletadas em visitas domiciliares e inquérito telefônico. Os indivíduos foram classificados segundo IMC em duas categorias: com e sem EP. Os fatores de risco cardiometabólico avaliados foram: circunferência de cintura aumentada, pressão arterial sistólica ou diastólica elevada, dislipidemia, glicemia de jejum ou hemoglobina glicada elevadas e resistência à insulina. Foram avaliados a frequência de consumo e o tamanho das porções de 30 grupos de alimentos e a qualidade da dieta, utilizando o Índice de Qualidade da Dieta Revisado para a população brasileira (IQD-R) e o Alternate Healthy Eating Index-2010 (AHEI) americano. Variáveis categóricas foram descritas e comparadas por teste qui-quadrado de Pearson e intervalo de confiança de 95%. Foram utilizados modelos de regressão logística e logística ordinal para avaliar a associação das variáveis investigadas com EP e com fatores de risco cardiometabólico. Resultados: No período avaliado, a prevalência de obesidade na população dobrou: de 10% (8,0; 12,5) em 2003 para 19,2% (17,8; 20,6) em 2015, sendo o maior aumento entre adolescentes e adultos do sexo feminino. Adultos, pessoas com maior renda e ex-fumantes apresentaram maior chance de ter EP, o qual aumentou 31% entre 2003 e 2008 e 126% entre 2003 e 2015, quando metade da população passou a ter EP. Se esse padrão se mantiver, espera-se que 77% da população tenha EP em 2030. Neste período, houve uma variação no tamanho das porções consumidas, com aumento em alguns grupos de alimentos (carne branca, salgados, café/chá, ovos) e redução em outros (arroz, carne vermelha, doces, massas, sanduiches, frios). Além disso, a frequência de consumo de seis grupos aumentou no período (arroz, carne branca, doces, frutas, sucos industrializados, torradas/biscoitos). Nesta população, o EP se associou positivamente com maiores porções de frios, petiscos fritos, sucos naturais e industrializados, pizza, carne vermelha, arroz, salgados, refrigerantes, sopas e açúcar. Em relação à qualidade da dieta, adolescentes com EP apresentaram pior qualidade da dieta para os componentes grãos integrais (IQD-R) e bebidas açucaradas (AHEI) em relação aos adolescentes sem EP. O IQD-R foi inversamente associado ao EP (OR=0.87; 0.80, 0.95) e ao número de fatores de risco cardiovascular (OR=0.89; 0.80, 0.98) nesta população. Conclusões: O EP aumentou substancialmente em um curto período na cidade de São Paulo, especificamente em alguns subgrupos e se associou positivamente com maior tamanho da porção de 11 grupos de alimentos na população geral e com menor qualidade da dieta entre os adolescentes. Adolescentes com menor qualidade da dieta também apresentaram maior chance de ter mais fatores de risco cardiometabólico. Esses achados podem subsidiar intervenções e políticas para redução da tendência de aumento do excesso de peso e melhora no consumo alimentar na população. / Introduction: Excess body weight (EBW, including overweight and obesity) is a major public health problem in the country. Understanding nutritional status over time and associated factors, such as diet, besides its consequences, such as cardiometabolic risk factors, is essential for the development of interventions in health. Objective: To evaluate the prevalence of overweight and obesity in the population of the city of São Paulo, as well as changes in diet between 2003 and 2015, and its association with cardiometabolic risk factors. Methods: We used data from the cross-sectional population based study ISA-Capital 2003, 2008, and 2015, with probabilistic sampling of residents of the city of São Paulo aged 12 years and more, of both sexes, totaling 8,682 individuals. Socioeconomic, anthropometric, and lifestyle data and 24-hour dietary recalls were collected in the households and by telephone. The individuals were categorized according to BMI in two categories: with and without EBW. The evaluated cardiometabolic risk factors were: central obesity, hypertension, dyslipidemia, high fasting glycaemia or glycated hemoglobin, and insulin resistance. We evaluated the frequency of intake and the portion sizes of 30 food groups, besides the diet quality, using the Revised Diet Quality Index for Brazilian population (BHEI-R) and the American Healthy Eating Index-2010 (AHEI). Categorical variables were described and compared by Pearson\'s chi-square test and 95% confidence interval. Ordinal logistic and logistic regression models were performed to evaluate the association of the investigated variables with EBW and cardiometabolic risk factors. Results: The prevalence of obesity in the population doubled in the evaluated period, from 10% (8.0; 12.5) in 2003 to 19.2% (17.8; 20.6) in 2015, with the highest increase among female adolescents and adults. Adults, those with higher income, and former smokers were more likely to have EBW, which increased 31% between 2003 and 2008, and 126% between 2003 and 2015, when half of the population had EBW. If this pattern continues, it is expected that 77% of the population will have EBW by 2030. During this period, there was a variation in the portion sizes of food groups consumed, with increase in some groups (white meat, salted snacks, coffee/tea, eggs) and decrease in others (rice, red meat, sweets, pasta, sandwiches, cold cuts). In addition, there was an increase in the frequency of consumption of six groups (rice, white meat, sweets, fruits, commercial juices, toasts/biscuits). In this population, EBW was positively associated with larger portions of cold cuts, fried snacks, fruit and commercial juices, pizza, red meat, rice, salted snacks, soft drinks, soups and sugar. Regarding diet quality, adolescents with EBW had lower diet quality for the components whole grains (BHEI-R) and sugar sweetened beverages (AHEI) compared with adolescents without EBW. BHEI-R was inversely associated with EBW (OR = 0.87, 0.80, 0.95) and cardiovascular risk factors count (OR = 0.89, 0.80, 0.98) in this population. Conclusions: EBW increased substantially over a short period in the city of São Paulo, especially in some subgroups and it was positively associated with larger portion size of 11 food groups in the population, and with lower diet quality among adolescents. Also, adolescents with lower diet quality were more likely to have higher number of cardiovascular risk factors. These findings can subsidize interventions and policies to reduce the tendency of increase in EBW and to improve food intake in the population.

Introdução: A avaliação do consumo alimentar de crianças e adolescentes representa um desafio na pesquisa epidemiológica. O Recordatório de 24 horas (R24h) é considerado um ótimo método para coletar dados dietéticos e é muito utilizado no acompanhamento alimentar da população. Porém requer entrevistadores altamente treinados, o que dificulta o uso em pesquisas de grandes amostras. Objetivo. Desenvolver um Recordatório de 24 horas estruturado e computadorizado (R24ec) para a avaliação do consumo alimentar de escolares e testar a usabilidade deste sistema. Métodos. Os alimentos escolhidos para o R24ec foram pré-definidos a partir de um banco de dados secundário. O R24ec inclui seis refeições prédefinidas, embutidas em questões que contextualizam as rotinas e locais do dia anterior. Contém ainda uma lista de alimentos dividida em grupos alimentares e subgrupos. Durante a entrevista, itens não listados podem ser adicionados em espaço previsto para este uso. A base de dados de composição de alimentos é composta pela tabela TACO. O R24ec ainda inclui um atlas fotográfico, contendo diferentes tipos de alimentos que serve como auxílio visual para que os entrevistados possam selecionar os tamanhos de porções. Realizou-se um teste com o R24ec em uma escola pública de Guarulhos com 73 escolares. A usabilidade foi testada por 17 profissionais da área da tecnologia da informação através de um questionário que avalia a usabilidade de sistemas utilizando a lógica Fuzzy baseado na ISO. Resultados. Foi desenvolvido um instrumento computadorizado contendo uma lista de aproximadamente 120 alimentos com média de três tamanhos de porções. Dos escolares pesquisados, cerca de 45 por cento fizeram alguma observação referente ao sistema durante o preenchimento do R24ec. 8,2 por cento tiveram problemas em relação à manipulação do sistema e 30 por cento tiveram problemas no preenchimento devido à incompreensão das perguntas, devido à idade ou maturidade. 67 por cento não cometeram erros ao preencher o R24ec. O teste de usabilidade apresentou para as métricas Controle de Erros, Eficiência e Satisfação amplitude Média (2), representando maior grau de confiança no resultado obtido. As métricas Facilidade de Aprender, Facilidade de Relembrar e Eficácia apresentaram amplitude Alta (3), o que representa um grau de confiança baixo. Conclusão. O Recordatório de 24 horas estruturado foi desenvolvido com sucesso e será hospedado no sitio da Universidade de São Paulo. O teste na escola com estudantes de 5ª a 8ª série apontou uma boa interação entre usuários e o sistema. O estudo de usabilidade com profissionais teve boas avaliações, porém deveria ser refeito com uma população maior para obterse maior confiança nos resultados / Introduction: The assessment of children and teenagers food intake brings a new challenge in epidemiological research. The 24-hour Dietary Recall (R24h) is considered an efficient method to collect diet data and is also very much used in population food monitoring. However, it requires highly qualified interviewers and this fact therefore generates great difficulty to the use of great sample research. Objectives: It aims at developing a structured and automated 24-hour Dietary Recall (R24ec) to evaluate school children food intake and test the usability of this system. Methods: The foods chosen to R24ec were pre-defined from a secondary database. The R24ec includes six pre-defined meals, incorporated in questions that contextualize the routines and places of the day before. It also contains a list of food divided in food groups and subgroups. During the interview, non-listed items can be included in proper blanks for this purpose. The database of food combination is composed of the TACO chart. The R24ec also includes a food photographic atlas, containing different types of food, used as visual help so that the interviewees can select the portion sizes. A test with school children was carried out with R24ec in a public school of Guarulhos. The usability was tested by 17 Information Technology professionals through a questionnaire that assesses the usability of systems using the Fuzzy logic, based on ISO. Results: An automated tool was developed, containing a list of approximately 120 types of food with an average of three portion sizes. Among the school children evaluated, around 45 per cent made some observation referring to the system during the R24ec filling process. 8,2 per cent had problems to manipulate the system and 30 per cent had problems to fill out the R24ec due to question misunderstandings, fact that can be related to the low level of maturity of the candidates. 67 per cent did not make any mistakes filling out the R24ec. The usability test presented to the metric Error Control, Efficiency and Satisfaction an average amplitude (2), representing a higher level of reliability in the obtained results. The metrics Facility to Learn, Facility to Remember and Efficiency presented high amplitude (3), which represents a low level of reliability. Conclusion: The structured 24-hour Dietary Recall was successfully developed and it will be posted in the Universidade de São Paulo website. The test with students from the Elementary School revealed good interaction between users and system. The usability study with professionals obtained good reviews, but it should be reapplied to a bigger audience in order to obtain a higher level of reliability towards the results

Objetivo: Identificar os determinantes da desnutrição infantil, na Comuna de Bom Jesus, em Angola. Métodos: Estudo transversal de base populacional com 742 crianças menores de 5 anos, Bom Jesus, Angola, 2010. Foi aplicado um inquérito por entrevista para coleta das variáveis independentes, seguido da aferição do peso e estatura. A classificação nutricional seguiu critérios da Organização Mundial da Saúde (2006). Para identificar os fatores associados aos déficits de peso/idade (P/I), estatura/idade (E/I) e peso/estatura (P/E), foram calculadas razões de prevalência mediante regressão de Poisson com variância robusta, utilizando-se modelos de determinantes hierarquizados. Resultados/Conclusões: Bairro de moradia, idade da criança e sexo associaram-se aos déficits nutricionais. No nível distal, observou-se maior chance entre crianças de domicílios sem energia elétrica, que viviam com o pai e cujo pai tinha outra família, respectivamente, aos déficits de P/I, P/E e E/I. No nível intermediário, maior chance foi observada entre crianças cujo domicílio era abastecido com água de rio ou lago e os déficits de P/I (RP:2,01; IC95%1,01-4,02) e E/I (RP:1,67; IC95%1,04-2,69), assim como entre déficit de P/E e crianças cuja ordem de nascimento era a partir do terceiro. Neste nível, observou-se menor chance do déficit de P/E entre crianças cuja idade da mãe era entre 25 e 34 anos e déficit de P/I e aquelas com um irmão menor de 5 anos. No nível proximal, maior chance foi observada em crianças que apresentaram vômito, 15 dias anteriores à pesquisa, e os déficits de P/I (RP:2,01; IC95%1,33-3,30) e E/I (RP:1,53; IC95%1,04-2,69). Maior chance foi observada entre o déficit linear e a manifestação de diarreia com muco e sangue, à véspera do inquérito. Esses resultados justificam a elaboração de intervenções voltadas à adequada assistência pré-natal e puericultura e melhoria do acesso à água de qualidade. / Objective: To identify the determinants of childhood malnutrition in the community of Bom Jesus in Angola. Methods: Population based cross-sectional study of 742 children under 5 in Bom Jesus, Angola, 2010. Data on the independent variables were collected using an interview, and height and weight were measured. Nutritional status was classified according to World Health Organization (2006) criteria. Poisson regression with robust variance was used to calculate prevalence ratios in order to identify factors associated with weight/age (W/A), height/age (H/A) and weight/height (W/H) deficiencies, using hierarchical models of the determinants. Results/Conclusions: Neighborhood of residence and age and sex of the child were associated with nutritional deficiencies. On the distal level, there was association between not being connected to an electricity grid, the presence of the father in the household and having another family being, respectively, to W/A. W/H and H/A deficiencies. At the intermediate level, there was association between household water supply from the river or lake and W/A (PR:2.01; 95%CI 1.01-4.02) and H/A (PR:1.67; 95%CI 1.04-2.69) deficiencies, as well as between W/H deficit and children whose birth order was 3rd or later. At this level there was an inverse association between W/H and mother being aged between 25 and 34 and W/A deficiency and having one sibling aged under 5. At the proximal level, was observed association between children who had presented vomiting within the 15 days preceding the research and W/A (PR:2.01; 95%CI 1.33-3.30) and H/A (PR:1.53; 95%CI 1.04-2.69) deficiencies. Association was observed between linear deficiency and diarrhea containing blood or mucus on the day before the survey. These results justify the creation of interventions directed at adequate pre-natal care and childcare, improving and increasing access to clean water and sanitation.

Introdução - A qualidade da dieta é um fator chave para as diferentes formas de má nutrição. No entanto, há ainda uma lacuna na coleta de dados necessários aos diagnósticos populacionais e vigilância. O Guia Alimentar para a População Brasileira (2014) aponta como componentes- chave da dieta a diversidade de alimentos in natura e minimamente processados e a restrição de alimentos ultraprocessados, com base na classificação NOVA dos alimentos. Objetivo - Desenvolver, adaptar e validar instrumentos de avaliação de duas dimensões da qualidade das dietas no Brasil. Métodos - Para a adaptação do instrumento de diversidade alimentar a partir do MDD-W (Diversidade Alimentar Mínima para Mulheres) da Organização das Nações Unidas para Alimentação e Agricultura (FAO) e desenvolvimento do instrumento de avaliação do consumo de alimentos ultraprocessados, foram analisados os dados do módulo de Consumo Alimentar coletado pela Pesquisa de Orçamentos Familiares (POF) do Instituto Brasileiro de Geografia e Estatística (IBGE), realizado entre maio de 2008 e maio de 2009 no Brasil. Foram selecionados 10 subgrupos de alimentos in natura e minimamente processados (grupo 1 classificação NOVA) e 10 subgrupos de alimentos ultraprocessados (grupo 4 classificação NOVA). Os instrumentos foram aplicados em amostra de conveniência (n=150) por meio de entrevistas a voluntários adultos da comunidade da Faculdade de Saúde Pública da Universidade de São Paulo, entre os meses de março e outubro de 2018. Em cada entrevista, após a aplicação de dois questionários simplificados, foi aplicado o método padrão ouro de recordatório alimentar completo de 24 horas. Foram gerados escores de diversidade e de ultraprocessados. Para o instrumento de diversidade, foi realizado cálculo de Kappa para análise de concordância, Alpha de Cronbach para análise de confiabilidade e análise fatorial confirmatória para validade de constructo. Para validação preditiva, as informações de composição nutricional foram avaliadas e calculadas com betas ajustados para regressão. Para o instrumento de consumo de alimentos ultraprocessados, foi avaliada a validade preditiva por meio do cálculo do Kappa ponderado e ajustado para prevalência e viés (PABAK), além da análise do percentual calórico de açúcar na dieta em relação ao escore. Resultados - Os valores de Kappa de concordância das respostas para o instrumento de diversidade alimentar variaram de 0,52 a 0,81. Sua confiabilidade obtida por meio do Alpha de Cronbach foi de 0,60, enquanto a análise fatorial confirmatória apontou bons índices de adequação, apresentando propriedades psicométricas satisfatórias. Para o instrumento de alimentos ultraprocessados, o valor de Kappa variou de 0,60 a 0,81, enquanto o valor de PABAK entre os quintis das variáveis obtido foi de 0,72, considerado substancial. Foi observada associação linear positiva significativa entre o os intervalos de escore de ultraprocessados e o percentual calórico de açúcar na dieta. Conclusões - Os instrumentos simplificados de diversidade alimentar e de consumo de alimentos ultraprocessados apresentam graus de validade satisfatórios. Esses instrumentos serão úteis para a pesquisa sobre qualidade da dieta em diferentes contextos de monitoramento, diagnósticos populacionais e estudos de intervenção e concepção de políticas públicas de saúde no Brasil. / Introduction - The diet quality is a key-factor for different types of malnutrition. However, there is a gap in the data collection in this area concerning population diagnosis and its monitoring. The Dietary Guidelines for the Brazilian Population (2014) points out that the keyfactors to diet quality are diversified fresh and minimally processed foods, and ultraprocessed foods restriction, in agreement with the NOVA classification. Objective - Development, adaptation and validation of tools to assess two dimensions of diet quality in Brazil. Methods - Adaptation of the Minimum Dietary Diversity for Women (MDD-W) indicator of the Food and Agriculture Organization (FAO). Development of a tool for assessing the consumption of ultra-processed foods. These both tools were based on data from the Food Consumption module, collected by the Household Budget Survey (POF) of the Brazilian Institute of Geography and Statistics (IBGE), which was conducted between May 2008 and May 2009 in Brazil. There was a selection of 10 subgroups of fresh and minimally processed foods (Group 1 according to NOVA classification) and 10 subgroups of ultraprocessed foods (Group 4 according to NOVA classification). The instruments were applied in a convenience sample (n=150) through interviews conducted between March and October 2018 with adult volunteers from the community of The Faculty of Public Health of The University of São Paulo. After the application of two simplified questionnaires at each interview, the gold standard method of 24-hour complete dietary recall was applied, from which diversity and ultra-processed scores were generated. For the diversity instrument, Kappa calculation was performed; Cronbach\'s Alpha was calculated for the reliability analysis; confirmatory factor analysis was applied in order to test the construct validity. For the predictive validation, the nutritional content was assessed and calculated on the grounds of betas adjusted values for regression. For the ultra-processed food consumption instrument, the predictive validity was assessed by calculating the adjusted for prevalence and bias (PABAK) Kappa, as well as sugar caloric percentage. Results - Kappa values of agreement of responses for the dietary diversity instrument ranged from 0.52 to 0.81. Its reliability, obtained by Cronbach\'s Alpha, was 0.60, while confirmatory factor analysis showed good adequacy indices and presented satisfactory psychometric properties. For the ultra-processed foods instrument, the Kappa value ranged from 0.60 to 0.81, while the PABAK value obtained among the quintiles of the variables was 0.72, which is considered substantial. Also, a positive linear association was found between sugar caloric percentage and ultra-processed food score intervals. Conclusions - The simplified instruments of dietary diversity and ultraprocessed food consumption presented satisfactory validity. These instruments can be useful for research on diet quality in different monitoring contexts and population diagnoses, as well as for intervention studies and designing of public health policies in Brazil.

This thesis examines the trajectory of food politics and its relationship with modernizing food landscapes. Today’s food politics discussions in the US and other high-income countries lean towards concerns with social issues, while in the recent past it was more focused on producing enough food to feed the growing population – two different situations with important concerns. The question this thesis explores is whether the modernization of society has given us a higher level of wellbeing and a better world through food landscapes and food policy, and how we came up with the current food situations we face today. It looks into the food landscape and policies of the US and its implications on wellbeing, and the dietary transition of Japan, comparing it to other modernized countries. It is important to look at these factors of food politics to see how other countries may handle a food transition, and whether or not there is a direction all countries can go in that will better benefit its food landscapes. Concerns not only lie in producing sufficient yields but also in wellbeing, cultural values, and true consumer demands. Only when true consumer demands are acknowledged individually can the trajectory of food politics be better harnessed and guided to a more desirable outcome than it has been going in so far.

Serum vitamin D deficiency/insufficiency, Chronic Kidney Disease (CKD) and elevated blood pressure are important health concerns especially among minorities with type 2 diabetes. The effect of vitamin D3 supplementation (cholecalciferol) at 6,000 IU/day (d) vs. 4,000 IU/d on kidney function and cardiovascular disease markers among Hispanics and African Americans with type 2 diabetes and hypovitaminosis D (/ml) was evaluated. Subjects (n=63) were recruited from two clinics in Miami-Dade County, FL. Fasting venous blood and fresh, single-voided first morning urine samples were collected from each participant by a certified phlebotomist and analyzed by Solstas Lab Partners, Davie, FL. Linear mixed models were used to compare the interaction between time and intervention. Least Significant Difference (LSD) comparisons were used to detect significant differences within and between 4,000 IU/d and 6,000 IU/d groups from baseline, 3 and 6 months. In the 4,000 IU/d and 6,000 IU/d groups, a significant increase in serum 25-hydroxy vitamin D [25(OH)D] levels were observed from baseline [(19.9±1.1 ng/mL) and (21.4±1.3 ng/mL)] to 3 months [(36.1±2.2 ng/mL, p3 longer than 6 months may be needed to determine sustained long term effects in kidney and cardiovascular disease markers. Further research could provide more information for translation of these findings into recommendations for individuals with CKD, hypertension and type 2 diabetes. The efficacy of vitamin D3 supplementation as complementary therapy for CKD and blood pressure in minority and other ethnic groups needs further investigation in larger and longer duration randomized controlled trials.

Introdução: A avaliação do consumo alimentar de crianças e adolescentes representa um desafio na pesquisa epidemiológica. O Recordatório de 24 horas (R24h) é considerado um ótimo método para coletar dados dietéticos e é muito utilizado no acompanhamento alimentar da população. Porém requer entrevistadores altamente treinados, o que dificulta o uso em pesquisas de grandes amostras. Objetivo. Desenvolver um Recordatório de 24 horas estruturado e computadorizado (R24ec) para a avaliação do consumo alimentar de escolares e testar a usabilidade deste sistema. Métodos. Os alimentos escolhidos para o R24ec foram pré-definidos a partir de um banco de dados secundário. O R24ec inclui seis refeições prédefinidas, embutidas em questões que contextualizam as rotinas e locais do dia anterior. Contém ainda uma lista de alimentos dividida em grupos alimentares e subgrupos. Durante a entrevista, itens não listados podem ser adicionados em espaço previsto para este uso. A base de dados de composição de alimentos é composta pela tabela TACO. O R24ec ainda inclui um atlas fotográfico, contendo diferentes tipos de alimentos que serve como auxílio visual para que os entrevistados possam selecionar os tamanhos de porções. Realizou-se um teste com o R24ec em uma escola pública de Guarulhos com 73 escolares. A usabilidade foi testada por 17 profissionais da área da tecnologia da informação através de um questionário que avalia a usabilidade de sistemas utilizando a lógica Fuzzy baseado na ISO. Resultados. Foi desenvolvido um instrumento computadorizado contendo uma lista de aproximadamente 120 alimentos com média de três tamanhos de porções. Dos escolares pesquisados, cerca de 45 por cento fizeram alguma observação referente ao sistema durante o preenchimento do R24ec. 8,2 por cento tiveram problemas em relação à manipulação do sistema e 30 por cento tiveram problemas no preenchimento devido à incompreensão das perguntas, devido à idade ou maturidade. 67 por cento não cometeram erros ao preencher o R24ec. O teste de usabilidade apresentou para as métricas Controle de Erros, Eficiência e Satisfação amplitude Média (2), representando maior grau de confiança no resultado obtido. As métricas Facilidade de Aprender, Facilidade de Relembrar e Eficácia apresentaram amplitude Alta (3), o que representa um grau de confiança baixo. Conclusão. O Recordatório de 24 horas estruturado foi desenvolvido com sucesso e será hospedado no sitio da Universidade de São Paulo. O teste na escola com estudantes de 5ª a 8ª série apontou uma boa interação entre usuários e o sistema. O estudo de usabilidade com profissionais teve boas avaliações, porém deveria ser refeito com uma população maior para obterse maior confiança nos resultados / Introduction: The assessment of children and teenagers food intake brings a new challenge in epidemiological research. The 24-hour Dietary Recall (R24h) is considered an efficient method to collect diet data and is also very much used in population food monitoring. However, it requires highly qualified interviewers and this fact therefore generates great difficulty to the use of great sample research. Objectives: It aims at developing a structured and automated 24-hour Dietary Recall (R24ec) to evaluate school children food intake and test the usability of this system. Methods: The foods chosen to R24ec were pre-defined from a secondary database. The R24ec includes six pre-defined meals, incorporated in questions that contextualize the routines and places of the day before. It also contains a list of food divided in food groups and subgroups. During the interview, non-listed items can be included in proper blanks for this purpose. The database of food combination is composed of the TACO chart. The R24ec also includes a food photographic atlas, containing different types of food, used as visual help so that the interviewees can select the portion sizes. A test with school children was carried out with R24ec in a public school of Guarulhos. The usability was tested by 17 Information Technology professionals through a questionnaire that assesses the usability of systems using the Fuzzy logic, based on ISO. Results: An automated tool was developed, containing a list of approximately 120 types of food with an average of three portion sizes. Among the school children evaluated, around 45 per cent made some observation referring to the system during the R24ec filling process. 8,2 per cent had problems to manipulate the system and 30 per cent had problems to fill out the R24ec due to question misunderstandings, fact that can be related to the low level of maturity of the candidates. 67 per cent did not make any mistakes filling out the R24ec. The usability test presented to the metric Error Control, Efficiency and Satisfaction an average amplitude (2), representing a higher level of reliability in the obtained results. The metrics Facility to Learn, Facility to Remember and Efficiency presented high amplitude (3), which represents a low level of reliability. Conclusion: The structured 24-hour Dietary Recall was successfully developed and it will be posted in the Universidade de São Paulo website. The test with students from the Elementary School revealed good interaction between users and system. The usability study with professionals obtained good reviews, but it should be reapplied to a bigger audience in order to obtain a higher level of reliability towards the results

Internalization and trafficking of ligand-receptor complex rely on a particular set of proteins, e.g. small GTPase protein Rab5 and its activators called guanine nucleotide exchange factors. Rab5-activating protein 6 (RAP6), a Vps9-containing protein, may participate in Rab5-mediated insulin signaling and receptor trafficking. A dicarbonyl compound methylglyoxal was found to alter insulin signaling in preadipocytes. This dissertation aimed to investigate the association of RAP6 activity on 3T3-L1 preadipocyte differentiation and those driven by methylglyoxal. Overexpression of RAP6 inhibited preadipocyte differentiation, Ser473-phosphorylation of Akt1, and expression of adipogenic marker PPARγ, but not C/EBPα. Methylglyoxal (10 µM) increased preadipocyte differentiation, proliferation and expression of PPARγ, C/EBPα and p-Akt1-Ser473, but appeared to be neutralized by RAP6 overexpression. The findings suggest that RAP6 may be a key modulator in regulating the stimulatory effect of methylglyoxal on preadipocyte differentiation. The associations of predominant methylglyoxal-derived adduct, methylglyoxal hydroimidazolone 1 (MGH1), with selected risk factors of chronic diseases in Black participants with and without type 2 diabetes (n=234 controls and n=254 cases) were also investigated. Only in individuals with diabetes, MGH1 levels were positively associated with fasting plasma glucose (B=0.240, p=0.037), homocysteine (B=0.355, p=0.014) and triglyceride (B=0.190, p=0.049). Being African Americans with type 2 diabetes was associated with lower MGH1 levels as compared to being Haitian American with diabetes (B=-0.334, p=0.016). The findings suggest that methylglyoxal may be linked to hyperglycemia and metabolic changes in type 2 diabetes, and may differently impact the development of diabetes across Black subgroups.

McMaster University MASTER OF PUBLIC HEALTH (2020) Hamilton, Ontario (Health Research Methods, Evidence, and Impact) TITLE: Applying Machine Learning to Determine Nutrients Predictive of Cardiovascular Disease Using Canadian Linked Population-Based Data AUTHOR: Jason D. Morgenstern, B.Sc. (University of Guelph), M.D. (Western University) SUPERVISOR: Professor L.N. Anderson, NUMBER OF PAGES: xv, 121 / The use of big data and machine learning may help to address some challenges in nutritional epidemiology. The first objective of this thesis was to explore the use of machine learning prediction models in a hypothesis-generating approach to evaluate how detailed dietary features contribute to CVD risk prediction. The second objective was to assess the predictive performance of the models. A population-based retrospective cohort study was conducted using linked Canadian data from 2004 – 2018. Study participants were adults age 20 and older (n=12 130 ) who completed the 2004 Canadian Community Health Survey, Cycle 2.2, Nutrition (CCHS 2.2). Statistics Canada has linked the CCHS 2.2 data to the Discharge Abstracts Database and the Canadian Vital Statistics Death database, which were used to determine cardiovascular outcomes (stroke or ischemic heart disease events or deaths). Conditional inference forests were used to develop models. Then, permutation feature importance (PFI) and accumulated local effects (ALEs) were calculated to explore contributions of nutrients to predicted disease. Supplement-use (median PFI (M)=4.09 x 10-4, IQR=8.25 x 10-7 – 1.11 x 10-3) and caffeine (M=2.79 x 10-4, IQR= -9.11 x 10-5 – 5.86 x 10-4) had the highest median PFIs for nutrition-related features. Supplement-use was associated with decreased predicted risk of CVD (accumulated local effects range (ALER)= -3.02 x 10-4 – 2.76 x 10-4) and caffeine was associated with increased predicted risk (ALER= -9.96 x 10-4 – 0.035). The best-performing model had a logarithmic loss of 0.248. Overall, many non-linear relationships were observed, including threshold, j-shaped, and u-shaped. The results of this exploratory study suggest that applying machine learning to the nutritional epidemiology of CVD, particularly using big datasets, may help elucidate risks and improve predictive models. Given the limited application thus far, work such as this could lead to improvements in public health recommendations and policy related to dietary behaviours. / Thesis / Master of Public Health (MPH) / This work explores the potential for machine learning to improve the study of diet and disease. In chapter 2, opportunities are identified for big data to make diet easier to measure. Also, we highlight how machine learning could find new, complex relationships between diet and disease. In chapter 3, we apply a machine learning algorithm, called conditional inference forests, to a unique Canadian dataset to predict whether people developed strokes or heart attacks. This dataset included responses to a health survey conducted in 2004, where participants’ responses have been linked to administrative databases that record when people go to hospital or die up until 2017. Using these techniques, we identified aspects of nutrition that predicted disease, including caffeine, alcohol, and supplement-use. This work suggests that machine learning may be helpful in our attempts to understand the relationships between diet and health.

Household-level characteristics have been shown to be associated with food insecurity but studies among vulnerable populations are sparse. A food security assessment was developed to determine food security and collect sociodemographic and household level data across San Luis Obispo County. The assessments were administered to vulnerable groups through interviews at multiple sites across the County. Three household characteristics (marital status, number of children in the household and number of workers in the household) were examined in this analysis. A total of 808 surveys were collected, 69% in English and 31% in Spanish. Through ethnicity-stratified sequentially adjusted logistic regression models, the association between food insecurity and household characteristics were tested, controlling for sociodemographic, economic and other potentially mediating variables. In the fully adjusted model for Hispanic/Latino households, associations were observed with number of children in the household and workers in the household, but confidence intervals were wide. In the fully adjusted model for White households, marital status was weakly associated with food insecurity. In both groups, per capita monthly income was strongly associated with food insecurity. Several interrelated household and individual level variables determined a households food security status. Because of this complexity, comprehensive social and economic changes are needed to improve food security in California and the rest of the United States. Also, different processes associated with race/ethnicity and coping strategies with regard to food insecurity should be considered when designing studies, planning policies, and conducting interventions.