Identifying biologically implausible values in big longitudinal data: an example applied to child growth data from the Brazilian food and nutrition surveillance system.

BACKGROUND: Several strategies for identifying biologically implausible values in longitudinal anthropometric data have recently been proposed, but the suitability of these strategies for large population datasets needs to be better understood. This study evaluated the impact of removing population outliers and the additional value of identifying and removing longitudinal outliers on the trajectories of length/height and weight and on the prevalence of child growth indicators in a large longitudinal dataset of child growth data. METHODS: Length/height and weight measurements of children aged 0 to 59 months from the Brazilian Food and Nutrition Surveillance System were analyzed. Population outliers were identified using z-scores from the World Health Organization (WHO) growth charts. After identifying and removing population outliers, residuals from linear mixed-effects models were used to flag longitudinal outliers. The following cutoffs for residuals were tested to flag those: -3/+3, -4/+4, -5/+5, -6/+6. The selected child growth indicators included length/height-for-age z-scores and weight-for-age z-scores, classified according to the WHO charts. RESULTS: The dataset included 50,154,738 records from 10,775,496 children. Boys and girls had 5.74% and 5.31% of length/height and 5.19% and 4.74% of weight values flagged as population outliers, respectively. After removing those, the percentage of longitudinal outliers varied from 0.02% (<-6/>+6) to 1.47% (<-3/>+3) for length/height and from 0.07 to 1.44% for weight in boys. In girls, the percentage of longitudinal outliers varied from 0.01 to 1.50% for length/height and from 0.08 to 1.45% for weight. The initial removal of population outliers played the most substantial role in the growth trajectories as it was the first step in the cleaning process, while the additional removal of longitudinal outliers had lower influence on those, regardless of the cutoff adopted. The prevalence of the selected indicators were also affected by both population and longitudinal (to a lesser extent) outliers. CONCLUSIONS: Although both population and longitudinal outliers can detect biologically implausible values in child growth data, removing population outliers seemed more relevant in this large administrative dataset, especially in calculating summary statistics. However, both types of outliers need to be identified and removed for the proper evaluation of trajectories.

Analysis of the quality of prenatal data of pregnant women attended at Healthcare Services in the city of São Paulo between 2012 and 2020.

OBJECTIVE: To analyze the quality of data collected during prenatal care recorded in the Integrated Health Care Management System (SIGA) of the Municipal Department of Health of São Paulo from 2012 to 2020. METHODS: Descriptive study using SIGA data and the variables: maternal height (cm), weight (kg) measured throughout pregnancy, gestational age at prenatal consultation, systolic (SBP) and diastolic (DBP) blood pressure (in mmHg), and body mass index (BMI) at the beginning of pregnancy (up to 8 weeks). Quality analysis was carried out by calculating the indicators: percentage of incompleteness and zero values of all variables studied, percentage of implausible values for height, weight, BMI; preference for terminal digit of weight and height, and normality of distributions. RESULTS: The database of pregnant women made available for analysis included 8,046,608 records and 1,174,115 women. The percentage of incompleteness and zero values was low (<1%) in all original variables of the system. There are more records at the end of pregnancy. For the four original variables of interest in the database (weight, height, SBP, DBP), there is a clear preference for the terminal digit. The variables of interest did not present an approximately normal distribution during the evaluated period. CONCLUSION: The quality analysis showed the need for improving the standardization of information collection and recording, the rounding of measurements and the need for encouraging pregnant women to start prenatal care as soon as possible, in such a way that it is important to invest in data quality, through educational resources for professionals who work in health care.

Quality of child anthropometric data from SISVAN, Brazil, 2008-2017.

OBJECTIVE: To evaluate the quality of anthropometric data of children recorded in the Food and Nutrition Surveillance System (SISVAN) from 2008 to 2017. METHOD: Descriptive study on the quality of anthropometric data of children under five years of age admitted in primary care services of the Unified Health System, from the individual databases of SISVAN. Data quality was annually assessed using the indicators: coverage, completeness, sex ratio, age distribution, weight and height digit preference, implausible z-score values, standard deviation, and normality of z-scores. RESULTS: In total, 73,745,023 records and 29,852,480 children were identified. Coverage increased from 17.7% in 2008 to 45.4% in 2017. Completeness of birth date, weight, and height corresponded to almost 100% in all years. The sex ratio was balanced and approximately similar to the expected ratio, ranging from 0.8 to 1. The age distribution revealed higher percentages of registrations from the ages of two to four years until mid-2015. A preference for terminal digits "zero" and "five" was identified among weight and height records. The percentages of implausible z-scores exceeded 1% for all anthropometric indices, with values decreasing from 2014 onwards. A high dispersion of z-scores, including standard deviations between 1.2 and 1.6, was identified mainly in the indices including height and in the records of children under two years of age and residents in the North, Northeast, and Midwest regions. The distribution of z-scores was symmetric for all indices and platykurtic for height/age and weight/age. CONCLUSIONS: The quality of SISVAN anthropometric data for children under five years of age has improved substantially between 2008 and 2017. Some indicators require attention, particularly for height measurements, whose quality was lower especially among groups more vulnerable to nutritional problems.

Maternal and child nutrition programme of investigation within the 100 Million Brazilian Cohort: study protocol.

INTRODUCTION: There is a limited understanding of the early nutrition and pregnancy determinants of short-term and long-term maternal and child health in ethnically diverse and socioeconomically vulnerable populations within low-income and middle-income countries. This investigation programme aims to: (1) describe maternal weight trajectories throughout the life course; (2) describe child weight, height and body mass index (BMI) trajectories; (3) create and validate models to predict childhood obesity at 5 years of age; (4) estimate the effects of prepregnancy BMI, gestational weight gain (GWG) and maternal weight trajectories on adverse maternal and neonatal outcomes and child growth trajectories; (5) estimate the effects of prepregnancy BMI, GWG, maternal weight and interpregnancy BMI changes on maternal and child outcomes in the subsequent pregnancy; and (6) estimate the effects of maternal food consumption and infant feeding practices on child nutritional status and growth trajectories. METHODS AND ANALYSIS: Linked data from four different Brazilian databases will be used: the 100 Million Brazilian Cohort, the Live Births Information System, the Mortality Information System and the Food and Nutrition Surveillance System. To analyse trajectories, latent-growth, superimposition by translation and rotation and broken stick models will be used. To create prediction models for childhood obesity, machine learning techniques will be applied. For the association between the selected exposure and outcomes variables, generalised linear models will be considered. Directed acyclic graphs will be constructed to identify potential confounders for each analysis investigating potential causal relationships. ETHICS AND DISSEMINATION: This protocol was approved by the Research Ethics Committees of the authors' institutions. The linkage will be carried out in a secure environment. After the linkage, the data will be de-identified, and pre-authorised researchers will access the data set via a virtual private network connection. Results will be reported in open-access journals and disseminated to policymakers and the broader public.

Gestational weight gain according to the Brazilian charts and its association with maternal and infant adverse outcomes.

BACKGROUND: The lack of gestational weight gain (GWG) recommendations for low- and middle-income countries is a significant concern. OBJECTIVES: To identify the ranges on the Brazilian GWG charts associated with lowest risks of selected adverse maternal and infant outcomes. METHODS: Data from 3 large Brazilian datasets were used. Pregnant individuals aged ≥18, without hypertensive disorders or gestational diabetes were included. Total GWG was standardized to gestational age-specific z-scores according to Brazilian GWG charts. A composite infant outcome was defined as the occurrence of any of small-for-gestationa lage (SGA), large-forgestationa lage (LGA), or preterm birth. In a separate sample, postpartum weight retention (PPWR) was measured at 6 and/or 12 mo postpartum. Multiple logistic and Poisson regressions were performed with GWG z-scores as the exposure and individual and composite outcomes. GWG ranges associated with the lowest risk of the composite infant outcome were identified using noninferiority margins. RESULTS: For the neonatal outcomes, 9500 individuals were included in the sample. For PPWR, 2602 and 7859 individuals were included at 6 and 12 mo postpartum, respectively. Overall, 7.5% of the neonates were SGA, 17.6% LGA, and 10.5% were preterm. Higher GWG z-scores were positively associated with LGA birth, whereas lower z-scores were positively associated with SGA births. The risk of the selected adverse neonatal outcomes were lowest (within 10% of lowest observed risk) when individuals with underweight, normal weight, overweight, or obesity gained between 8.8-12.6; 8.7-12.4; 7.0-8.9; and 5.0-7.2 kg, respectively. These gains correspond to probabilities of PPWR ≥5 kg at 12 mo of 30% for individuals with under and normal weight, and <20% for overweight and obesity. CONCLUSIONS: This study provided evidence to inform new GWG recommendations in Brazil.

Analysis of the quality of prenatal data of pregnant women attended at Healthcare Services in the city of São Paulo between 2012 and 2020 / Análise da qualidade dos dados de pré-natal de gestantes atendidas em Unidades Básicas de Saúde do município de São Paulo entre 2012 e 2020

ABSTRACT Objective: To analyze the quality of data collected during prenatal care recorded in the Integrated Health Care Management System (SIGA) of the Municipal Department of Health of São Paulo from 2012 to 2020. Methods: Descriptive study using SIGA data and the variables: maternal height (cm), weight (kg) measured throughout pregnancy, gestational age at prenatal consultation, systolic (SBP) and diastolic (DBP) blood pressure (in mmHg), and body mass index (BMI) at the beginning of pregnancy (up to 8 weeks). Quality analysis was carried out by calculating the indicators: percentage of incompleteness and zero values of all variables studied, percentage of implausible values for height, weight, BMI; preference for terminal digit of weight and height, and normality of distributions. Results: The database of pregnant women made available for analysis included 8,046,608 records and 1,174,115 women. The percentage of incompleteness and zero values was low (<1%) in all original variables of the system. There are more records at the end of pregnancy. For the four original variables of interest in the database (weight, height, SBP, DBP), there is a clear preference for the terminal digit. The variables of interest did not present an approximately normal distribution during the evaluated period. Conclusion: The quality analysis showed the need for improving the standardization of information collection and recording, the rounding of measurements and the need for encouraging pregnant women to start prenatal care as soon as possible, in such a way that it is important to invest in data quality, through educational resources for professionals who work in health care.

RESUMO Objetivo: Analisar a qualidade dos dados coletados no acompanhamento pré-natal registrados no Sistema Integrado de Gestão da Assistência à Saúde (SIGA) da Secretaria Municipal de Saúde de São Paulo de 2012 a 2020. Métodos: Estudo descritivo utilizando dados do SIGA e as variáveis: altura materna (cm), peso (kg) medido ao longo da gestação, idade gestacional na consulta pré-natal, pressão arterial (em mmHg) sistólica (PAS) e diastólica (PAD), e índice de massa corporal (IMC) no início da gestação (até 8 semanas). A análise da qualidade foi realizada por meio do cálculo dos indicadores: percentual de incompletude e valores zero de todas as variáveis estudadas, percentual de valores implausíveis de estatura, peso, IMC; preferência por dígito terminal do peso e estatura, e normalidade das distribuições. Resultados: Base de dados de gestantes disponibilizada para análise incluía 8.046.608 registros e 1.174.115 mulheres. O percentual de incompletude e valores zeros foi baixo (<1%) em todas as variáveis originais do sistema. Existe maior número de registros ao final da gestação. Para as quatro variáveis de interesse originais do banco de dados (peso, altura, PAS, PAD), existe clara preferência por dígito terminal. As variáveis de interesse não apresentaram distribuição aproximadamente normal durante o período avaliado. Conclusão: A análise da qualidade mostrou necessidade de melhoria na padronização da coleta e do registro das informações, no arredondamento das medidas e na necessidade de incentivar as gestantes a iniciar o pré-natal o quanto antes; por isso, é importante investir na qualidade do dado, por meio de recursos educativos para profissionais que atuam na assistência.

Quality of child anthropometric data from SISVAN, Brazil, 2008-2017 / Qualidade dos dados antropométricos infantis do Sisvan, Brasil, 2008-2017

ABSTRACT OBJECTIVE To evaluate the quality of anthropometric data of children recorded in the Food and Nutrition Surveillance System (SISVAN) from 2008 to 2017. METHOD Descriptive study on the quality of anthropometric data of children under five years of age admitted in primary care services of the Unified Health System, from the individual databases of SISVAN. Data quality was annually assessed using the indicators: coverage, completeness, sex ratio, age distribution, weight and height digit preference, implausible z-score values, standard deviation, and normality of z-scores. RESULTS In total, 73,745,023 records and 29,852,480 children were identified. Coverage increased from 17.7% in 2008 to 45.4% in 2017. Completeness of birth date, weight, and height corresponded to almost 100% in all years. The sex ratio was balanced and approximately similar to the expected ratio, ranging from 0.8 to 1. The age distribution revealed higher percentages of registrations from the ages of two to four years until mid-2015. A preference for terminal digits "zero" and "five" was identified among weight and height records. The percentages of implausible z-scores exceeded 1% for all anthropometric indices, with values decreasing from 2014 onwards. A high dispersion of z-scores, including standard deviations between 1.2 and 1.6, was identified mainly in the indices including height and in the records of children under two years of age and residents in the North, Northeast, and Midwest regions. The distribution of z-scores was symmetric for all indices and platykurtic for height/age and weight/age. CONCLUSIONS The quality of SISVAN anthropometric data for children under five years of age has improved substantially between 2008 and 2017. Some indicators require attention, particularly for height measurements, whose quality was lower especially among groups more vulnerable to nutritional problems.

RESUMO OBJETIVOS Avaliar a qualidade dos dados antropométricos de crianças registradas no Sistema de Vigilância Alimentar e Nutricional (Sisvan) no período 2008-2017. MÉTODOS Estudo descritivo sobre a qualidade dos dados antropométricos de crianças menores de 5 anos atendidas nos serviços de atenção primária do Sistema Único de Saúde, a partir das bases de dados individuais do Sisvan. A qualidade dos dados foi avaliada anualmente por meio dos indicadores: cobertura, completude, razão entre sexos, distribuição da idade, preferência por dígitos de peso e estatura, valores de escore-z implausíveis, desvio-padrão e normalidade dos escores-z. RESULTADOS No total, 73.745.023 registros e 29.852.480 crianças foram identificados. A cobertura aumentou de 17,7% em 2008 para 45,4% em 2017. A completude da data de nascimento, peso e estatura correspondeu a quase 100% para todos os anos. A razão entre sexos foi equilibrada e aproximadamente similar a razão esperada, variando entre 0,8 e 1. A distribuição da idade revelou maiores percentuais de registros entre as idades de 2 a 4 anos até meados de 2015. Uma preferência pelos dígitos terminais "zero" e "cinco" foi identificada entre os registros de peso e estatura. As porcentagens de escores-z implausíveis excederam 1% para todos os índices antropométricos, com redução dos valores a partir de 2014. Uma alta dispersão dos escores-z, incluindo desvios-padrão entre 1,2 e 1,6, foi identificada principalmente nos índices incluindo estatura e nos registros de crianças menores de 2 anos e residentes das regiões Norte, Nordeste e Centro-Oeste. A distribuição dos escores-z foi simétrica para todos os índices e platicúrtica para estatura/idade e peso/idade. CONCLUSÕES A qualidade dos dados antropométricos do Sisvan para crianças menores de 5 anos melhorou substancialmente entre 2008 e 2017. Alguns indicadores requerem atenção, sobretudo para medidas de estatura, cuja qualidade foi principalmente inferior entre os grupos mais vulneráveis a agravos nutricionais.

Prevalence and temporal trends in prepregnancy nutritional status and gestational weight gain of adult women followed in the Brazilian Food and Nutrition Surveillance System from 2008 to 2018.

Prepregnancy body mass index (BMI) and gestational weight gain (GWG) are the most investigated indicators of maternal nutritional status, which is a modifiable factor that plays a vital role in maternal and infant health. This study describes prepregnancy BMI and GWG of 840,243 women with 2,087,765 weight observations in the Brazilian Food and Nutrition Surveillance System from 2008 to 2018. Prepregnancy BMI was classified according to the World Health Organization cut-offs. Total GWG was calculated from weight measurements taken after 36 weeks of pregnancy and classified according to the Institute of Medicine guidelines. Temporal trends in prepregnancy BMI status were examined, and maps were used to evaluate changes in excessive GWG in each Brazilian federation unit. On overall, prepregnancy overweight and obesity increased from 22.6% to 28.8% and from 9.8% to 19.8%, respectively, between 2008 and 2018. The prevalence of excessive GWG rose from 34.2% to 38.7% during the same period and in 11 of the 27 Brazilian federation units between 2008 and 2016. Women with underweight showed the highest values for mean total GWG for all the compared years (overall variation from 12.3 to 13.1 kg), followed by those with normal weight (11.9 to 12.5 kg), overweight (10.1 to 10.9 kg) and obesity (from 8.2 to 8.9 kg). Within each BMI group, values remained fairly stable throughout the studied period for first- and second-trimester GWG and total GWG. These results help to fill a significant gap in understanding the distribution of prepregnancy BMI and GWG in Brazilian women.

Association of ADIPOQ-rs2241766 and FTO-rs9939609 genetic variants with body mass index trajectory in women of reproductive age over 6 years of follow-up: the PREDI study.

BACKGROUND: Interindividual variations in body mass index (BMI) can be partially explained by genetic differences. We aimed to examine the association of the ADIPOQ-rs2241766, LEP-rs7799039 and FTO-rs9939609 genetic variants with BMI trajectory in women of reproductive age over 6 years of follow-up. METHODS: This was a prospective study that used data from 435 women of the PREDI Study conducted in Brazil. Socioeconomic, biological and anthropometric data were collected at four time points: 2012 (baseline) in the maternity hospital, and 2013-14, 2016-17 and 2018 (1st, 2nd and 3rd follow-ups) at the participant's home. Genotyping was performed by PCR-RFLP. Linear mixed-effect and Poisson regression models were used to address the association of ADIPOQ, LEP and FTO genotypes with BMI and overweight/obesity status. RESULTS: Women carrying the risk allele (TA or AA) of the FTO-rs9939609 genetic variant had a 1.16 kg/m2 higher BMI over the follow-up period than those carrying the wild-type genotype (TT), even when adjusted for potential confounders (95% CI: 0.23-2.10, p = 0.015). The risk of obesity associated with the FTO-TA or AA genotype decreased over the years, demonstrating an influence of time on its trajectory (IRR = 0.99, 95% CI: 0.98-0.99, p = 0.016). There was no variation in BMI trajectories for the ADIPOQ-rs2241766, LEP-rs7799039 or FTO-rs9939609 genetic variant. CONCLUSIONS: The results of this study suggest that monitoring women of reproductive age with ADIPOQ-rs2241766 TG/GG or FTO-rs9939609 TA/AA genotypes may be an important strategy to reduce maternal excess body weight and, consequently, the long-term public health burden of obesity.

Maternal mental health and gestational weight gain in a Brazilian Cohort.

Depression and anxiety are common during pregnancy, but little is known about the influence of these disorders on gestational weight gain (GWG). Data from a prospective cohort of pregnant women followed in a public healthcare center in Rio de Janeiro, Brazil, were used to evaluate the association of depression, anxiety, and suicide risk with GWG. GWG was evaluated at 5-13, 20-26, 30-36, and 37-42 weeks, and GWG adequacy was determined. Statistical analyses included linear mixed-effect models and Poisson regression. We evaluated 206 women, in which 15% (n = 31) presented major depressive disorder, 19.4% (n = 34) suicide risk and 10% (n = 21) generalized anxiety disorder at baseline. Women with depression at the first trimester, persistent depressive symptoms, and anxiety symptoms at the second trimester presented significantly lower rates of GWG per week compared to those without depression or anxiety, respectively. Persistent depressive symptoms represented a 2.40 (95% CI 1.20; 4.81; p = 0.013) increase in the risk of insufficient GWG. There was no significant association between generalized anxiety disorder or suicide risk with GWG. The presence of depression, depressive symptoms, and anxiety during pregnancy were associated with lower GWG rates. Persistent depressive symptoms during pregnancy were directly associated with insufficient GWG.

Vitamin D during pregnancy and its association with birth outcomes: a Brazilian cohort study.

BACKGROUND/OBJECTIVES: Evidence is lacking on how variation in vitamin D concentrations during gestation affects perinatal outcomes. Therefore, we evaluated the association between vitamin D serum concentrations during pregnancy and neonatal outcomes. METHODS: A prospective cohort of 180 healthy Brazilian pregnant women was followed and plasma 25-hydroxyvitamin [25(OH)D, nmol/L] was measured at 5-13 (baseline), 20-26 and 30-36 gestational weeks. Birth weight (BW), birth length (BL), BW z-scores, BL z-scores, first minute Apgar, small for gestational age (SGA), large for gestational age (LGA) and preterm birth were the outcomes. Multiple linear and Poisson regression models were estimated. Best linear unbiased prediction of random coefficients model was used to determine the association between the mean rate of change in vitamin D during pregnancy concentrations and neonatal outcomes. RESULTS: Mean (SD) BW was 3300 (600) g, BW z-score 0.34 (1.11), BL 49.3 (3.3) cm, BL z-score 0.44 (1.5), and first minute Apgar score 8.2 (1.4). Prevalence of SGA, LGA and preterm birth were 6%, 18% and 13%, respectively. 25(OH)D was directly associated with the risk of preterm birth at all trimesters. Incidence-rate ratios were 1.02, 1.05 and 1.04 for the 1st, 2nd and 3rd trimester, respectively. Mean rate of change during pregnancy in 25(OH)D was directly associated with BW z-score (ß: 0.36, 95% CI 0.07; 0.65), LGA risk (IRR: 1.97, 95% CI 1.07; 3.63) and preterm birth (IRR: 7.35, 95% CI 2.99; 18.07). CONCLUSIONS: Mean 25(OH)D rate of change during pregnancy was directly associated with BW z-scores, and increased LGA and preterm birth risk.

Brazilian Maternal and Child Nutrition Consortium: establishment, data harmonization and basic characteristics.

Pooled data analysis in the field of maternal and child nutrition rarely incorporates data from low- and middle-income countries and existing studies lack a description of the methods used to harmonize the data and to assess heterogeneity. We describe the creation of the Brazilian Maternal and Child Nutrition Consortium dataset, from multiple pooled longitudinal studies, having gestational weight gain (GWG) as an example. Investigators of the eligible studies published from 1990 to 2018 were invited to participate. We conducted consistency analysis, identified outliers, and assessed heterogeneity for GWG. Outliers identification considered the longitudinal nature of the data. Heterogeneity was performed adjusting multilevel models. We identified 68 studies and invited 59 for this initiative. Data from 29 studies were received, 21 were retained for analysis, resulting in a final sample of 17,344 women with 72,616 weight measurements. Fewer than 1% of all weight measurements were flagged as outliers. Women with pre-pregnancy obesity had lower values for GWG throughout pregnancy. GWG, birth length and weight were similar across the studies and remarkably similar to a Brazilian nationwide study. Pooled data analyses can increase the potential of addressing important questions regarding maternal and child health, especially in countries where research investment is limited.

Association between vitamin D status during pregnancy and total gestational weight gain and postpartum weight retention: a prospective cohort.

OBJECTIVES: To evaluate the association between vitamin D status during pregnancy and total gestational weight gain (GWG), GWG rates and postpartum weight retention. METHODS: Prospective cohort of 163 women from Rio de Janeiro was followed at 5th-13th (baseline), 20th-26th, 30th-36th gestational weeks and at 30-62 days postpartum. Plasma 25-hydroxyvitamin D [25(OH)D] was evaluated during each trimester and was categorized as adequate (≥50 nmol/L) or inadequate (<50 nmol/L). GWG (kg) was calculated as the difference between the weight measured at baseline and 36th-42th gestational weeks. GWG rates (kg/week) were calculated between each visit. Postpartum weight retention (kg) was analysed as the difference between weights measured at 30-62 days postpartum and 5th-13th gestational weeks. Statistical analyses were performed using linear regression models that included interaction terms between vitamin D status and first trimester body mass index (BMI) (<25/≥25 kg/m2). Confounders were selected based on a directed acyclic graph. RESULTS: The prevalence of vitamin D inadequacy was 16.6%, 9.9% and 10.6% in the first, second and third trimesters, respectively. Overweight women with vitamin D inadequacy in the first (ß = 3.70; 95% CI 0.09; 7.31, p-value = 0.045) and third trimester (ß = 4.59, 95% CI 0.07; 9.10, p-value = 0.047) presented higher increases in total GWG than did women with vitamin D adequacy. This association was also observed between first trimester vitamin D status and GWG rates between visits 1 and 2 (ß = 0.17; 95% CI 0.13; 0.36, p-value = 0.07). CONCLUSIONS: There was an interaction effect of first trimester BMI (≥25 kg/m2) on the association between first and third trimester vitamin D status and GWG.