https://tedebc.ufma.br/jspui/handle/tede/882 2026-08-31T01:08:05Z https://tedebc.ufma.br/jspui/handle/tede/7189 Título: PREDIÇÃO DE NASCIMENTO PRÉ-TERMO EXTREMO EM DUAS CIDADES BRASILEIRAS Autor: BRITO, Emmanuelle Novaes de Vasconcelos Primeiro orientador: THOMAZ, Erika Bárbara Abreu Fonseca Abstract: Introduction: Extreme preterm birth (EPTB) is a major public health problem associated with high neonatal and infant morbidity and mortality. Its occurrence is influenced by multiple clinical, sociodemographic, and behavioral factors, making prevention particularly challenging. Although relatively uncommon, EPTB has substantial consequences for maternal and child health. In this context, machine learning (ML) techniques offer a promising approach for identifying pregnant women at increased risk of EPTB. Objective: To develop and evaluate machine learning-based predictive models to identify predictors of extreme preterm birth among live births from two Brazilian cities. Methods: This analytical cross-sectional study was nested within the 2010 BRISA prenatal and birth cohorts conducted in São Luís, Maranhão, and Ribeirão Preto, São Paulo, Brazil. The inclusion of these cohorts allowed the investigation of maternal and child health in distinct socioeconomic and demographic settings. The study included singleton pregnancies receiving care in public and private healthcare services, comprising 6,414 mother–child pairs from São Luís and 8,273 from Ribeirão Preto, totaling 14,687 observations. A sensitivity analysis was performed using the variable threatened preterm labor during the current pregnancy, which represents ongoing pathophysiological changes potentially related to the onset of the outcome and is considered an important predictor. Six analytical scenarios were developed, corresponding to the combined cohort and the São Luís and Ribeirão Preto cohorts, each analyzed with and without this variable. Based on expert knowledge and previous literature, 119 variables common to both cohorts, 121 variables in São Luís, and 119 variables in Ribeirão Preto were selected as potential predictors, including socioeconomic, demographic, behavioral, reproductive, and maternal healthcare characteristics. The BORUTA feature selection algorithm identified 50 relevant variables, from which the eight most important predictors were retained. Fourteen machine learning algorithms were evaluated in each scenario: penalized logistic regression (PLR), support vector machines (svmLinear and svmRadial), random forest (rf, ranger, and Rborist), boosting methods (glmboost and AdaBoost.M1), nearest neighbor methods (Knn e KKnn), artificial neural networks (nnet), tree-based partitioning models (cforest and ctree2) and generalized additive models (GAM). Model performance was assessed using accuracy, sensitivity, specificity, precision, F1-score, and the area under the receiver operating characteristic curve (AUC-ROC). Results: All analytical scenarios showed class imbalance between EPTB and non-EPTB births. Model performance varied across the six scenarios. When threatened preterm labor was included, kNN showed the best overall performance in the combined cohort (sensitivity = 0.774, AUC-ROC = 0.806, F1-score = 0.049), svmRadial performed best in São Luís (sensitivity = 0.653, AUC-ROC = 0.732, F1-score = 0.063), and ranger achieved the highest performance in Ribeirão Preto (AUC-ROC = 0.860, sensitivity = 0.771, F1-score = 0.053). When this variable was excluded, nnet performed best in the combined cohort (sensitivity = 0.561, AUC-ROC = 0.648, F1-score = 0.019), GAM in São Luís (sensitivity = 0.623, AUCROC = 0.587, F1-score = 0.024), and Rborist in Ribeirão Preto (AUC-ROC = 0.629, sensitivity = 0.573, F1-score = 0.016). Overall, kNN and nnet demonstrated the best predictive performance in scenarios with and without the threatened preterm labor variable, respectively. The eight most important predictors were threatened preterm labor during the current pregnancy, threatened miscarriage during the current pregnancy, maternal occupation, per capita income, alcohol consumption during pregnancy, self-reported maternal race/skin color, planned pregnancy, and hypertension diagnosed before or during pregnancy. Conclusion: Machine learning models can predict extreme preterm birth with good predictive performance using maternal health, fetal, socioeconomic, behavioral, and prenatal care variables. Early identification of pregnancies at increased risk may support timely interventions and improve maternal and neonatal outcomes. These findings provide evidence for the potential application of machine learning in prenatal risk stratification and encourage further validation in different populations and clinical settings. Instituição: Universidade Federal do Maranhão Tipo do documento: Dissertação 2026-05-26T00:00:00Z https://tedebc.ufma.br/jspui/handle/tede/7178 Título: Padrões espaço-temporais e sociodemográficos de agravos nutricionais e alimentares da população brasileira Autor: SILVEIRA, Victor Nogueira da Cruz Primeiro orientador: FRANÇA, Ana Karina Teixeira da Cunha Abstract: Nutritional issues, manifesting in both anthropometric and dietary terms, represent detri- mental conditions that impact the quality of life and future prospects of individuals and populations. These issues are typically intertwined, as inadequate dietary intake—specifically of industrial products such as ultra-processed foods (UPFs)—directly affects nutritional status. Furthermore, given their multifactorial nature, sociodemographic factors also in- fluence both aspects. Accordingly, this thesis analyzed the sociodemographic conditions directly influencing UPF consumption among the Brazilian population aged 10 years and older (Article 1) and identified combinations of effects that increased or decreased the con- sumption of UPFs and inadequate nutritional/metabolic profiles (INMP) among Brazilian adults (Article 2). Furthermore, the trend, progression and factors associated with hospi- talizations for mortality caused by malnutrition were analyzed (Article 3) and the spatial distribution of consumption of certain classes of UPF in four different age groups by Brazil- ian micro-regions (Article 4). A Bayesian hierarchical modeling approach was employed to assess the relationship between sociodemographic predictors and ultra-processed food consumption across these micro-regions. Variable selection was initially performed using the Spike-and-Slab technique, which allows for the probabilistic inclusion of covariates in the linear predictor. Spatial structure was analyzed by comparing five models (Null, Independent, CAR, BYM, and BYM2) using the Watanabe-Akaike Information Criterion (WAIC). The BYM2 model provided the best fit, allowing the residual variance to be de- composed into two components: a structured random effect capturing spatial dependence between neighbors, and an unstructured effect associated with specific heterogeneity, both weighted by a mixing parameter. Analyses were conducted in the R environment, utilizing the Rstan package for Markov Chain Monte Carlo (MCMC) sampling and R-INLA to cal- culate the spatial variance scaling factor. Chain convergence was verified via the Gelman- Rubin diagnostic and visual inspection of traceplots. This statistical framework enabled the control of spatial autocorrelation and the precise identification of sociodemographic de- terminants of food consumption. The results indicate a high prevalence of ultra-processed food (UPF) consumption in Brazil, particularly among the youth population (≤ 19 years old), who surpassed adults in all categories—most notably regarding sugar-sweetened bev- erages (SSB – 66.1%) and sweets/cookies (59.1%). Spatially, a higher probability of SSB and sweets consumption was observed in the Central-West and Southeast regions, whereas the consumption of sausages and savory snacks was concentrated along the Northeastern coast and in the South. Variable selection using the Spike-and-Slab method identified the *Bolsa Família* Program (PBF), illiteracy, and income as the primary predictors. In the final model (BYM2), micro-regions with higher PBF participation showed lower odds of SSB consumption among youth (OR = 0.86) and processed meat consumption among adults (OR = 0.93), but higher odds of savory snack consumption (OR = 1.11). Analysis of random effects confirmed strong spatial dependence (high λ), indicating that the geographic environment influences dietary patterns. Adults exhibited a more homoge- neous distribution of consumption across the territory, whereas high-consumption clusters among youth were more localized, reinforcing the need for existing regionally structured public policies aimed at controlling UPF consumption. The findings reveal a heteroge- neous distribution of ultra-processed food consumption in Brazil, with higher prevalence among children and adolescents—particularly regarding sugar-sweetened beverages and sweets. The impact of the *Bolsa Família* Program was twofold: it acted as a protec- tive factor regarding the consumption of beverages and sausages but was associated with increased savory snack consumption among adults. Spatial analysis confirmed regional patterns, with higher consumption concentrated in the South and Southeast. It is con- cluded that the socioeconomic environment and age shape dietary choices, necessitating regionalized public policies that strengthen nutrition education and access to unprocessed foods in order to mitigate health inequalities. Instituição: Universidade Federal do Maranhão Tipo do documento: Tese 2026-07-09T00:00:00Z https://tedebc.ufma.br/jspui/handle/tede/7170 Título: Experiências de pacientes, familiares e trabalhadores da saúde: desafios e aprendizagens para o cuidado em situações de crise Autor: OLIVEIRA, Marcela Lobão de Primeiro orientador: LAMY, Zeni Carvalho Abstract: INTRODUCTION: The COVID-19 pandemic produced profound social, economic, cultural, and political transformations, demanding rapid institutional adaptations and modifying behaviors, social relations, and forms of care. Its effects were felt unevenly in different countries, regions, and territories, highlighting structural vulnerabilities and distinct capacities to respond to the health crisis. In Brazil, Manaus experienced the collapse of its health system, marked by insufficient hospital beds, a shortage of professionals, and a lack of essential supplies such as oxygen, which led to a Cooperation Plan between states of the federation to enable the transfer of patients from Manaus to other states, with the goal of making it possible to assist hospitalized patients. OBJECTIVE: To analyze the experiences of patients, family members, and healthcare workers in seeking care in the context of the COVID-19 pandemic. METHODOLOGY: Qualitative research, grounded in hermeneutic phenomenology focused on experiences of illness, suffering, and therapeutic journey. Conducted between March 2022 and July 2023 with patients from Manaus who had been hospitalized at the University Hospital of the Federal University of Maranhão, their families, and healthcare workers responsible for their care in São Luís. Semi-structured interviews were conducted on the Google Meet platform, after the completion of the Free and Informed Consent Form (FICF) and sociodemographic questionnaire. Participants included 26 healthcare workers, 12 patients, and 10 family members. The reports were submitted to content analysis. RESULTS: Presented in the form of two scientific articles. The first, “Paths and Obstacles in the Journey for Healthcare During the Covid-19 Pandemic: Manaus as an emblematic case,” analyzed the perspectives of patients and their families regarding therapeutic itineraries and experiences of illness in the context of a health collapse. Different strategies for seeking care were mobilized from the perception of the first signs of illness. Access to health services in Manaus was marked by fear of contamination, insecurity, and the perception of health units as a “war zone.” Transfer to São Luís was understood as a survival strategy, and the reception and treatment received aroused gratitude. The return to Manaus was marked by ambivalent feelings – relief, fear, and uncertainty – highlighting that the journey during the pandemic was marked by tensions and displacements produced in the relationships between subjects, emotions, social contexts, and political decisions amidst a global health crisis. The second article, “Experiences of Death and Disruptions in the Covid-19 Pandemic: Reconfigurations of care, grief, and daily life,” analyzed the meanings attributed to the disruptions and suffering resulting from the experiences of death lived by patients, family members, and healthcare workers during the pandemic. The social management of death and funeral rites was reconfigured, producing lasting impacts on the ways of experiencing grief. The pandemic was understood as a disruptive event whose effects persist latently in the bodies, discourses, and silences of those who experienced it, aggravated by the invisibility of losses and the insufficiency of reparative policies focused on care and collective memory. FINAL CONSIDERATIONS: The COVID-19 pandemic profoundly transformed individual and collective experiences of illness, care, and death. The restrictions imposed to control the spread of the virus produced disruptions in daily life, intensifying experiences of suffering, insecurity, and vulnerability. The health collapse went beyond the biomedical dimension, affecting social relations, care practices, and ways of processing grief. The temporal distance between the end of the health emergency and the consolidation of international agreements aimed at preparing for future crises reveals the limits of multilateral cooperation in defense of life and human rights. At the same time, the naturalization and silencing of lived experiences contribute to the delegitimization of the suffering of those who were exposed to the precariousness of care and social protection systems during the pandemic. Instituição: Universidade Federal do Maranhão Tipo do documento: Tese 2026-06-24T00:00:00Z https://tedebc.ufma.br/jspui/handle/tede/7169 Título: Boas práticas de atenção ao parto na perspectiva das puérperas e dos trabalhadores de maternidades da rede cegonha: uma avaliação por meio da teoria de resposta ao item Autor: FIGUEIREDO, Kely Nayara dos Reis Silva Primeiro orientador: SANTOS, Alcione Miranda dos Abstract: Good labor and birth care practices are defined by actions to humanize the care for the parturient, child and family, which seek better maternal and perinatal care outcomes. The main objective of this thesis was to evaluate the good practices of care in labor and delivery in Brazil, based on the Item Response Theory (IRT), consisting of two articles. The first article defined a scale to measure the level of supply of good practices in labor and delivery, latent trait estimated by the IRT through the three-parameter-one-dimensional logistic model (M3PL). For that purpose, this study used data from 2427 health professionals which participated in the evaluative research “Evaluation of care during labor and birth in maternity hospitals in RC”. As a result, a scale with seven positioned items and three anchor levels was obtained. On the first level are the maternity hospitals that only offered the strategies defined for the reception and always encouraged the pregnant woman to walk around. The second level maternity hospitals, in addition to the items above, guaranteed the pregnant woman the right to freely choose a companion during hospitalization, offered massage and ball and different delivery positions. On the third level are the maternity hospitals which, in addition to the previous items, offered birth stools. It was possible to identify the contribution of each item in measuring the level of supply of good care practices for labor and delivery, allowing the construction of an interpretative scale for the evaluation of maternity hospitals in the Rede Cegonha at three levels. The second article evaluated good practices in labor and delivery in the perception of postpartum women in RC maternity hospitals and identified the factors associated with the offer of good practices. The sample consisted of 3,716 postpartum women participating in the evaluative research. The three-parameter-one-dimensional logistic model (M3PL) was used. The results showed that the items referring to reception better discriminate the puerperal women regarding the application of good practices, and indigenous puerperal women, women with less schooling, primiparous women and those coming from the North Region showed a lower level of adequacy to good practices in labor and delivery. The puerperal women assisted in highly complex maternity hospitals had greater access to good practices during labor and delivery. The methodology used in this study makes it a pioneer among national research. The advantages of using the Item Response Theory in the investigation of good labor and delivery practices were highlighted. Instituição: Universidade Federal do Maranhão Tipo do documento: Tese 2026-03-15T00:00:00Z