Comparing Methods for Identifying Biologically Implausible Values in Height, Weight, and Body Mass Index Among Youth

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.

Frequency of anthropometric implausible values estimated from different methodologies: a systematic review and meta-analysis

CONTEXT

Poor anthropometric data quality affect the prevalence of malnutrition and could harm public policy planning.

OBJECTIVE

This systematic review and meta-analysis was designed to identify different methods to evaluate and clean anthropometric data, and to calculate the frequency of implausible values for weight and height obtained from these methodologies.

DATA SOURCES

Studies about anthropometric data quality and/or anthropometric data cleaning were searched for in the MEDLINE, LILACS, SciELO, Embase, Scopus, Web of Science, and Google Scholar databases in October 2020 and updated in January 2023. In addition, references of included studies were searched for the identification of potentially eligible studies.

DATA EXTRACTION

Paired researchers selected studies, extracted data, and critically appraised the selected publications.

DATA ANALYSIS

Meta-analysis of the frequency of implausible values and 95% confidence interval (CI) was estimated. Heterogeneity (I2) and publication bias were examined by meta-regression and funnel plot, respectively.

RESULTS

In the qualitative synthesis, 123 reports from 104 studies were included, and in the quantitative synthesis, 23 studies of weight and 14 studies of height were included. The study reports were published between 1980 and 2022. The frequency of implausible values for weight was 0.55% (95%CI, 0.29-0.91) and for height was 1.20% (95%CI, 0.44-2.33). Heterogeneity was not affected by the methodological quality score of the studies and publication bias was discarded.

CONCLUSIONS

Height had twice the frequency of implausible values compared with weight. Using a set of indicators of quality to evaluate anthropometric data is better than using indicators singly.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO registration no. CRD42020208977.

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.

Prevalence of obesity and abdominal obesity and their association with metabolic-related conditions in Vietnamese adults: an analysis of Vietnam STEPS survey 2009 and 2015

Background

The abdominal obesity trends and prevalence are important contributing factors to significant rise of many noncommunicable diseases in Vietnam but have not been well-documented in the literature. This study aimed to describe the prevalence and trends of obesity and abdominal obesity in Vietnam from 2009 to 2015 and evaluate how different definitions of obesity and abdominal obesity are associated with metabolic-related conditions.

Methods

We conducted a secondary analysis based on the Vietnam STEPS (STEPwise approach to Surveillance) cross-sectional Survey 2009 and 2015. Obesity and abdominal obesity were defined using the body mass index (BMI), waist circumference (WC), and waist-hip ratio (WHR) cut-offs from the World Health Organization (WHO) and International Diabetes Federation (IDF).

Findings

Depending on the specific cut-offs, from 2009 to 2015, obesity prevalence increased from 0.8%-10% to 1.7%-16.4% in women and from 0.8%-10.3% to 1.7%-15% in men; abdominal obesity prevalence increased from 3%-31.3% to 8%-41.7% in women and from 0.3%-19.3% to 0.4%-25% in men. Abdominal obesity using WC-IDF and WHR-WHO definitions had noticeably higher sensitivity and lower specificity for metabolic-related conditions compared to the other four criteria. All anthropometric measurements were statistically correlated with biomarkers/blood pressure in 2009 and 2015 except for fasting glucose. Only WC-IDF and WHR-WHO definitions showed consistent association with all reported metabolic-related conditions regardless of sex and survey years.

Interpretation

The prevalence of obesity and abdominal obesity in Vietnam is increasing rapidly, especially abdominal obesity in women regardless of the criteria used. More studies are needed to investigate how using different diagnostic criteria for obesity and abdominal obesity could better identify metabolic-related conditions.

Funding

Authors received no funding for this study.

Data quality control in longitudinal epidemiologic studies: conditional studentized residuals from linear mixed effects models for outlier detection in the setting of pediatric chronic kidney disease

PURPOSE

Quality control in longitudinal cohort studies is critical for valid epidemiologic inference. Conditional studentized residuals (CSRs) derived from linear mixed effects models offer efficient individual-specific quality control. We present the utility of CSRs for outlier detection in an applied example using data from the Chronic Kidney Disease in Children cohort.

METHODS

Longitudinal linear mixed effects models with glomerular filtration rate (GFR) as the outcome were fit for observations prior to kidney replacement therapy, stratified by nonglomerular or glomerular diagnosis, and for a subset after receiving a kidney transplant. For each model, CSRs were calculated and values ≥±5 were considered potential outliers for further investigation.

RESULTS

A total of 1096 participants contributed 6881 annual measures of GFR across the two diagnostic groups and after transplant. In all models, the fixed effects captured progressive GFR decline. CSRs provided measures of individual-level deviations from the modeled trajectories (random + fixed effects) and were easily visualized in longitudinal plots. A total of 38 potential outliers from 32 participants were detected and further investigated for quality control.

CONCLUSIONS

This example demonstrated how longitudinal models can provide CSRs to detect individual-specific outliers. CSRs should be considered as part of quality control for longitudinal epidemiologic studies.

Genetic investigation of the contribution of body composition to anorexia nervosa in an electronic health record setting

Anorexia nervosa (AN) is a psychiatric disorder defined by anthropometric symptoms, such as low body weight, and cognitive-behavioral symptoms, such as restricted eating, fear of weight gain, and distorted body image. Recent studies have identified a genetic association between AN and metabolic/anthropometric factors, including body mass index (BMI). Although the reported associations may be under pleiotropic genetic influences, they may represent independent risk factors for AN. Here we examined the independent contributions of genetic predisposition to low body weight and polygenic risk (PRS) for AN in a clinical population (Vanderbilt University Medical Center biobank, BioVU). We fitted logistic and linear regression models in a retrospective case-control design (123 AN patients, 615 age-matched controls). We replicated the genetic correlations between PRS_BMI and AN (p = 1.12 × 10^-3, OR = 0.96), but this correlation disappeared when controlling for lowest BMI (p = 0.84, OR = 1.00). Additionally, we performed a phenome-wide association analysis of the PRS_AN and found that the associations with metabolic phenotypes were attenuated when controlling for PRS_BMI. These findings suggest that the genetic association between BMI and AN may be a consequence of the weight-related diagnostic criteria for AN and that genetically regulated anthropometric traits (like BMI) may be independent of AN psychopathology. If so, individuals with cognitive-behavioral symptomatology suggestive of AN, but with a higher PRS_BMI, may be under-diagnosed given current diagnostic criteria. Furthermore, PRS_BMI may serve as an independent risk factor for weight loss and weight gain during recovery.

Cleaning of anthropometric data from PCORnet electronic health records using automated algorithms

Objective

To demonstrate the utility of growthcleanr, an anthropometric data cleaning method designed for electronic health records (EHR).

Materials and Methods

We used all available pediatric and adult height and weight data from an ongoing observational study that includes EHR data from 15 healthcare systems and applied growthcleanr to identify outliers and errors and compared its performance in pediatric data with 2 other pediatric data cleaning methods: (1) conditional percentile (cp) and (2) PaEdiatric ANthropometric measurement Outlier Flagging pipeline (peanof).

Results

687 226 children (<20 years) and 3 267 293 adults contributed 71 246 369 weight and 51 525 487 height measurements. growthcleanr flagged 18% of pediatric and 12% of adult measurements for exclusion, mostly as carried-forward measures for pediatric data and duplicates for adult and pediatric data. After removing the flagged measurements, 0.5% and 0.6% of the pediatric heights and weights and 0.3% and 1.4% of the adult heights and weights, respectively, were biologically implausible according to the CDC and other established cut points. Compared with other pediatric cleaning methods, growthcleanr flagged the most measurements for exclusion; however, it did not flag some more extreme measurements. The prevalence of severe pediatric obesity was 9.0%, 9.2%, and 8.0% after cleaning by growthcleanr, cp, and peanof, respectively.

Conclusion

growthcleanr is useful for cleaning pediatric and adult height and weight data. It is the only method with the ability to clean adult data and identify carried-forward and duplicates, which are prevalent in EHR. Findings of this study can be used to improve the growthcleanr algorithm.

Novel Pediatric Height Outlier Detection Methodology for Electronic Health Records via Machine Learning With Monotonic Bayesian Additive Regression Trees

OBJECTIVE

To create a new methodology that has a single simple rule to identify height outliers in the electronic health records (EHR) of children.

METHODS

We constructed 2 independent cohorts of children 2 to 8 years old to train and validate a model predicting heights from age, gender, race and weight with monotonic Bayesian additive regression trees. The training cohort consisted of 1376 children where outliers were unknown. The testing cohort consisted of 318 patients that were manually reviewed retrospectively to identify height outliers.

RESULTS

The amount of variation explained in height values by our model, R2 , was 82.2% and 75.3% in the training and testing cohorts, respectively. The discriminatory ability to assess height outliers in the testing cohort as assessed by the area under the receiver operating characteristic curve was excellent, 0.841. Based on a relatively aggressive cutoff of 0.075, the outlier sensitivity is 0.713, the specificity 0.793; the positive predictive value 0.615 and the negative predictive value is 0.856.

CONCLUSIONS

We have developed a new reliable, largely automated, outlier detection method which is applicable to the identification of height outliers in the pediatric EHR. This methodology can be applied to assess the veracity of height measurements ensuring reliable indices of body proportionality such as body mass index.

Infant Body Mass Index or Weight-for-Length and Risk of Undernutrition in Childhood Among Children with Cystic Fibrosis

OBJECTIVES

To compare performance of weight-for-length and body mass index as estimators of undernutrition in children with cystic fibrosis (CF).

STUDY DESIGN

We analyzed pediatric anthropometric data from the Cystic Fibrosis Foundation Patient Registry. Undernutrition was defined by weight-for-length z score (WFLZ) or body mass index z score (BMIZ) ≤-1 (15th-percentile). Group 1, reference group, consisted of subjects with both BMIZ and WFLZ >-1; group 2: BMIZ ≤-1 and WFLZ >-1; group 3: BMIZ >-1 and WFLZ ≤-1; and group 4: BMIZ and WFLZ ≤-1. Group differences in length-for-age-Z across ages 2-24 months were tested using generalized estimating equations. The association of group at age 2 months with BMIZ <-1 at age 6 years was tested using logistic regression adjusted for demographic and disease characteristics.

RESULTS

Overall, 163 482 anthropometric measurements were available from 12 640 individuals, of whom 16.8% were discordant for undernutrition status at age 2 months. Discordance (1.5%-10%) was less common with increasing age. Length-for-age-Z was lower in group 2 than group 1 and group 3 between birth and 24 months (P < .05). Odds of WFLZ-defined undernourished at 2 months were lower for shorter individuals (OR 1.5, CI 1.4-1.6, P < .001). Undernutrition risk at age 6 years was greater for group 2 vs group 3 (OR 1.9 vs 1.0, P < .001).

CONCLUSIONS

Infants with cystic fibrosis classified as undernourished by BMIZ, but not WFLZ, had greater risk of undernourished status later in childhood. Infants with low BMIZ but normal WFLZ tended to be shorter, suggesting BMIZ may better capture undernourished status than WFLZ in shorter infants.

Association Between Length of Only-Child Period During Early Childhood and Overweight at Age 8-A Population-Based Longitudinal Study in Japan

INTRODUCTION

Prior studies have shown that children who are the only child are more likely to be overweight compared to their peers with siblings, regardless of whether they are the oldest, in the middle, or youngest. The study objective was to clarify whether there is an association between the length of the only-child period and the risk of overweight in firstborns who experienced an only-child period during early childhood before their siblings were born.

METHODS

A total of 7,576 first-born boys and 7,229 first-born girls were examined from a nationwide longitudinal survey in Japan. The length of the only-child period was determined by "birth interval"; i.e., the interval between the birth of the index child and the birth of the second child. It was categorized as short (<1.5 years), moderate (between 1.5 and 4 years), long (between 4 and 8 years), and only-child (the second baby was not born for 8 years). Overweight was defined as body mass index (BMI) z-score 1 standard deviation or more at age 8. Logistic regression was used to examine the association between length of only-child period and childhood overweight, adjusting for covariates.

RESULTS

Moderate birth interval was inversely associated with being overweight in comparison with only-child in both boys (odds ratio (OR): 0.83, 95% CI, 0.72-0.96) and girls (OR: 0.75, 95% CI, 0.63-0.88). Long birth interval also showed inverse association in boys (OR: 0.78, 95% CI, 0.62-0.97), and marginal inverse association in girls (OR: 0.80, 95% CI, 0.62-1.04).

CONCLUSION

First-born children who experienced short birth intervals did not show a different overweight risk from only-child. First-born children who experienced 1.5-8 years of the birth interval had a lower risk of childhood overweight compared with only-child.

Genetic potential and height velocity during childhood and adolescence do not fully account for shorter stature in cystic fibrosis

BACKGROUND

Despite improved health, shorter stature is common in cystic fibrosis (CF). We aimed to describe height velocity (HV) and contribution of height-related genetic variants to height (HT) in CF.

METHODS

HV cohort: standard deviation scores (-Z) for HT, mid-parental height-adjusted HT (MPAH), and HV were generated using our Pediatric Center's CF Foundation registry data. HV-Z was compared with population means at each age (5-17 y), the relationship of HV-Z with HT-Z assessed, and HT-Z compared with MPAH-Z. GRS cohort: HT genetic risk-Z (HT-GRS-Z) were determined for pancreatic exocrine sufficient (PS) and insufficient (PI) youth and adults from our CF center and their relationships with HT-Z assessed.

RESULTS

HV cohort: average HV-Z was normal across ages in our cohort but was 1.5× lower (p < 0.01) for each SD decrease in HT-Z. MPAH-Z was lower than HT-Z (p < 0.001). GRS cohort: HT-GRS-Z more strongly correlated with HT-Z and better explained height variance in PS (rho = 0.42; R²= 0.25) vs. PI (rho = 0.27; R² = 0.11).

CONCLUSIONS

Despite shorter stature compared with peers and mid-parental height, youth with CF generally have normal linear growth in mid- and late childhood. PI tempered the heritability of height. These results suggest that, in CF, final height is determined early in life in CF and genetic potential is attenuated by other factors.

IMPACT

Children with CF remain shorter than their healthy peers despite advances in care. Our study demonstrates that children with CF have persistent shorter stature from an early age and fail to reach their genetic potential despite height velocities comparable to those of average maturing healthy peers and similar enrichment in known height increasing single-nucleotide polymorphisms (SNPs). Genetic risk scores better explained variability in pancreatic sufficient than in pancreatic insufficient individuals, suggesting that other modifying factors are in play for pancreatic insufficient individuals with CF. Given the CF Foundation's recommendation to target not only normal body mass index, but normal height percentiles as well, this study adds valuable insight to this discussion.

Normalized measures and patient characteristics to identify undernutrition in infants and young children treated for cancer

BACKGROUND

Various measures and definitions for undernutrition are used in pediatrics. Younger children treated for cancer are at high risk, but lack well-defined risk-based screening and intervention.

METHODS

A retrospective study collected weight longitudinally for patients less than three years-old over two years after initiating cancer treatment. We included those diagnosed 2007-2015 at a large pediatric cancer center. Exclusion criteria included treatment starting outside our system, secondary or relapsed malignancy, or incomplete information. A decrease ≥1 in weight-for-age or weight-for-height z-score signified clinically significant weight loss. Univariate and multivariate models assessed hazards for developing first episode of clinically significant weight loss.

RESULTS

Of 372 patients, only 24.6% of patients lost 10% of weight, but 58.6% lost weight-for-age z-score ≥1 and 64.8% lost ≥1 weight-for-height z-score within two years of treatment initiation. Patients who lost weight were younger (median age 15 vs. 24 months, p < 0.001). Compared to patients diagnosed in the first year of life, those diagnosed 24-35 months were less likely to lose weight (HR 0.62, p < 0.001) and lost weight later (median time to weight loss 144 vs. 35 days). Higher treatment intensity increased weight loss risk (HR 2.30, p < 0.001) and decreased time to weight loss (35 vs. 154 days). No differences were found based on sex, diagnosis, enteral or parenteral nutrition, gastroenterology consults, or intensive care admissions.

CONCLUSIONS

Using normalized z-scores is more sensitive for identifying weight loss. Younger children are more likely to lose weight with higher intensity cancer therapy. Patient and treatment specific information should be used in risk stratifying weight loss screening and nutritional interventions.

Construction and Use of Body Weight Measures from Administrative Data in a Large National Health System: A Systematic Review

OBJECTIVE

Administrative data are increasingly used in research and evaluation yet lack standardized guidelines for constructing measures using these data. Body weight measures from administrative data serve critical functions of monitoring patient health, evaluating interventions, and informing research. This study aimed to describe the algorithms used by researchers to construct and use weight measures.

METHODS

A structured, systematic literature review of studies that constructed body weight measures from the Veterans Health Administration was conducted. Key information regarding time frames and time windows of data collection, measure calculations, data cleaning, treatment of missing and outlier weight values, and validation processes was collected.

RESULTS

We identified 39 studies out of 492 nonduplicated records for inclusion. Studies parameterized weight outcomes as change in weight from baseline to follow-up (62%), weight trajectory over time (21%), proportion of participants meeting weight threshold (46%), or multiple methods (28%). Most (90%) reported total time in follow-up and number of time points. Fewer reported time windows (54%), outlier values (51%), missing values (34%), or validation strategies (15%).

CONCLUSIONS

A high variability in the operationalization of weight measures was found. Improving methods to construct clinical measures will support transparency and replicability in approaches, guide interpretation of findings, and facilitate comparisons across studies.

Automated data cleaning of paediatric anthropometric data from longitudinal electronic health records: protocol and application to a large patient cohort

‘Big data’ in healthcare encompass measurements collated from multiple sources with various degrees of data quality. These data require quality control assessment to optimise quality for clinical management and for robust large-scale data analysis in healthcare research. Height and weight data represent one of the most abundantly recorded health statistics. The shift to electronic recording of anthropometric measurements in electronic healthcare records, has rapidly inflated the number of measurements. WHO guidelines inform removal of population-based extreme outliers but an absence of tools limits cleaning of longitudinal anthropometric measurements. We developed and optimised a protocol for cleaning paediatric height and weight data that incorporates outlier detection using robust linear regression methodology using a manually curated set of 6,279 patients’ longitudinal measurements. The protocol was then applied to a cohort of 200,000 patient records collected from 60,000 paediatric patients attending a regional teaching hospital in South England. WHO guidelines detected biologically implausible data in <1% of records. Additional error rates of 3% and 0.2% for height and weight respectively were detected using the protocol. Inflated error rates for height measurements were largely due to small but physiologically implausible decreases in height. Lowest error rates were observed when data was measured and digitally recorded by staff routinely required to do so. The protocol successfully automates the parsing of implausible and poor quality height and weight data from a voluminous longitudinal dataset and standardises the quality assessment of data for clinical and research applications.

Is it time to stop sweeping data cleaning under the carpet? A novel algorithm for outlier management in growth data

All data are prone to error and require data cleaning prior to analysis. An important example is longitudinal growth data, for which there are no universally agreed standard methods for identifying and removing implausible values and many existing methods have limitations that restrict their usage across different domains. A decision-making algorithm that modified or deleted growth measurements based on a combination of pre-defined cut-offs and logic rules was designed. Five data cleaning methods for growth were tested with and without the addition of the algorithm and applied to five different longitudinal growth datasets: four uncleaned canine weight or height datasets and one pre-cleaned human weight dataset with randomly simulated errors. Prior to the addition of the algorithm, data cleaning based on non-linear mixed effects models was the most effective in all datasets and had on average a minimum of 26.00% higher sensitivity and 0.12% higher specificity than other methods. Data cleaning methods using the algorithm had improved data preservation and were capable of correcting simulated errors according to the gold standard; returning a value to its original state prior to error simulation. The algorithm improved the performance of all data cleaning methods and increased the average sensitivity and specificity of the non-linear mixed effects model method by 7.68% and 0.42% respectively. Using non-linear mixed effects models combined with the algorithm to clean data allows individual growth trajectories to vary from the population by using repeated longitudinal measurements, identifies consecutive errors or those within the first data entry, avoids the requirement for a minimum number of data entries, preserves data where possible by correcting errors rather than deleting them and removes duplications intelligently. This algorithm is broadly applicable to data cleaning anthropometric data in different mammalian species and could be adapted for use in a range of other domains.

Association between grandparent coresidence and weight change among first-grade Japanese children

BACKGROUND

The prevalence of grandparent coresidence is increasing. However, the impact of grandparent coresidence on weight change among school-age children remains unclear.

OBJECTIVES

The objective of the study is to examine the association between grandparent coresidence and obesity-related behaviours and change in body mass index (BMI) z scores among school-age children.

METHODS

In total, 3422 caregivers of first-grade children in Adachi City, Tokyo, participated in surveys and health checkups in 2016 and 2017 with no change in their grandparent coresidence status (response rate: 80.1% and 81.4%, respectively). Association between grandparent coresidence and obesogenic dietary, physical activity, and screen-based sedentary behaviours was measured using Poisson regression with robust error variance analysis, while change in BMI z scores was determined by linear regression and adjusting for potential covariates.

RESULTS

Grandparent coresidence was associated with increased prevalence rate (PR) of irregular snack foods intake (PR: 1.38, 95% confidential interval (CI): 1.19, 1.61); however, children who live with grandparents showed a lower BMI z scores in the second grade (coefficient: -0.048, 95% CI: -0.094 to -0.0013) after adjustment for BMI z scores in the first grade, family sociodemographics, and obesogenic behaviours.

CONCLUSION

Grandparent coresidence is associated with lower BMI z scores among early primary school-age Japanese children living in urban areas.

Family food environment factors associated with obesity outcomes in early childhood

Background

In attempting to gain understanding of the family food environment (FFE), as a central context for the development of obesity and obesogenic eating behaviours during early childhood, attention has largely focused on the relationships of individual variables. This fails to capture the complex combinations of variables children are exposed to. To more authentically reflect the impact of the FFE on the development of obesity and obesogenic eating behaviours during early childhood, this study aims to derive composites of FFE variables using factor analysis.

Methods

FFE and eating behaviour data were available from 757 Australian children (2.0–5.0 years) via a parent-completed online survey. Children were categorised as normal weight, overweight or obese, based on parent-reported anthropometry (underweight children were excluded).

Results

Eight FFE factors were derived. Scores for factors ‘Negative Feeding Strategies’ and ‘Negative Nutrition Related Beliefs’ increased with child BMI category, while ‘Use of TV and devices’ and ‘Parent’s Nutrition Knowledge’ decreased. The FFE factor ‘Negative Feeding Strategies’ was positively associated with food fussiness, food responsiveness and slowness in eating, and negatively associated with parent body mass index (BMI) score. The FFE factor ‘Negative Nutrition Related Beliefs’ was positively associated with food responsiveness, as well as positively with parent BMI, male children, breastfeeding less than 6 months, and low-income status. The FFE factor ‘Television (TV) and devices’ was only positively associated with residing in a capital city. The FFE factor ‘Parent’s Nutrition Knowledge’ was negatively associated with slowness in eating, breastfeeding less than 6 months and low-income status, and positively with parent stress and residing in a capital city.

Conclusion

Consideration of the composite effect of FFE on child’s eating behaviours and obesity outcomes is important in guiding future research and obesity prevention initiatives by providing a more authentic picture of the FFE children are exposed to. Examining factors of FFE variables in conjunction with psycho-social variables, further articulates the reciprocal influence of these variables on environmental constructs thus assisting in understanding of inequitable distribution of obesity risk.

*keywords

childhood obesity, eating behaviours, early childhood, Family Food Environment, Factor Analysis,

The Protective Effect of Prenatal Social Support on Infant Adiposity in the First 18 Months of Life

OBJECTIVE

To determine whether prenatal social support was associated with infant adiposity in the first 18 months of life in a low-income, Hispanic sample, known to be at high risk of early child obesity.

STUDY DESIGN

We performed a longitudinal analysis of 262 low-income, Hispanic mother-infant pairs in the control group of the Starting Early child obesity prevention trial. Prenatal social support was measured using an item from the Maternal Social Support Index. We used multilevel modeling to predict weight-for-length z-score trajectories from birth to age 18 months and logistic regression to predict macrosomia and overweight status at ages 6, 12, and 18 months.

RESULTS

High prenatal social support was independently associated with lower infant adiposity trajectories from birth to age 18 months (B = -0.40; 95% CI, -0.63 to -0.16), a lower odds of macrosomia (aOR = 0.35; 95% CI, 0.15-0.80), and a lower odds of overweight at ages 12 (aOR = 0.28; 95% CI, 0.10-0.74) and 18 months (aOR = 0.35; 95% CI, 0.14-0.89). Prenatal social support was not significantly associated with overweight status at age 6 months.

CONCLUSIONS

Prenatal social support may protect against excessive infant adiposity and overweight in low-income, Hispanic families. Further research is needed to elucidate mechanisms underlying these associations and to inform preventive strategies beginning in pregnancy.

An examination of children's eating behaviours as mediators of the relationship between parents' feeding practices and early childhood body mass index z-scores

OBJECTIVE

Parent's use of restrictive feeding practices is associated with child weight. Similarly, the literature shows that children's eating behaviours are also associated with child weight. Given this interrelationship between children's eating behaviours, restrictive feeding practices and child weight, examination of possible mediator relationships is warranted. This study aimed to examine the relationships between overt restriction and covert restriction with child body mass index z-scores (BMIz) and determine if children's eating behaviours (satiety responsiveness and food responsiveness) act as mediators.

METHOD

Parents of Australian children (n = 977) 2.0-5.0 years of age (49.4% male) provided data in an online survey on child eating behaviours, parent's restrictive feeding practices and child anthropometrics (modified z-scores were created to screen for biologically implausible values). Correlation analysis was used to determine variables to include in mediation models. Hayes' PROCESS macros in spss was used to examine mediation, controlling for covariates of child BMIz.

RESULTS

Overt restriction was the only parent feeding practice related to child BMIz (B = 0.132, P = 0.04). Mediation analysis showed that the indirect effect of overt restriction on child BMIz (controlling for child age, gender, parent BMI and income) became non-significant when controlling for food responsiveness, thus suggesting full mediation, explaining 5.75% of the relation.

CONCLUSION

Overt restriction and covert restriction have distinctly different relationships with children's eating behaviours. Food responsiveness appears an important intermediary in the relationship between overt restriction and child BMIz.

Effect of a Breakfast in the Classroom Initiative on Obesity in Urban School-aged Children: A Cluster Randomized Clinical Trial

IMPORTANCE

Serving breakfast in the classroom is promoted to increase participation in the federal School Breakfast Program. However, little is known about the effect of breakfast in the classroom on children's weight status.

OBJECTIVE

To evaluate the effect of a breakfast in the classroom initiative, which combined breakfast in the classroom with breakfast-specific nutrition education, on overweight and obesity among urban children in low-income communities.

DESIGN, SETTING, AND PARTICIPANTS

A cluster-randomized clinical trial among 1362 fourth- through sixth-grade students from low-income urban communities across 2.5 years. Sixteen kindergarten through eighth grade Philadelphia public schools with universal breakfast participated. Participants were recruited in September 2013, and the intervention began in January 2014. Data analysis took place from April 1, 2018, to August 30, 2018.

INTERVENTIONS

Intervention schools received a program that included breakfast in the classroom and breakfast-specific nutrition education. Control schools continued breakfast before school in the cafeteria and standard nutrition education.

MAIN OUTCOMES AND MEASURES

The primary outcome was the combined incidence of overweight and obesity. Secondary outcomes included the combined prevalence of overweight and obesity, incidence and prevalence of obesity, changes in body mass index (BMI) z score, and School Breakfast Program participation.

RESULTS

Among the 1362 students, mean (SD) age was 10.8 (0.96) years and 700 (51.4%) were female; 907 (66.6%) were black, 233 (17.1%) were Hispanic, 100 (7.3%) were white, 83 (6.1%) were Asian, and 39 were of multiple or other race/ethnicity. After 2.5 years, students in intervention schools had participated in the School Breakfast Program 53.8% of days, compared with 24.9% of days among students in control schools (β = 0.33; 95% CI, 0.22-0.42). There was no difference between intervention and control schools in the combined incidence of overweight and obesity after 2.5 years (11.7% vs 9.3%; odds ratio [OR] 1.31; 95% CI, 0.85-2.02; P = .22). However, the incidence (11.6% vs 4.4%; OR, 2.43; 95% CI, 1.47-4.00) and prevalence (28.0% vs 21.2%; OR, 1.46; 95% CI, 1.11-1.92) of obesity were higher in intervention schools than in control schools after 2.5 years.

CONCLUSIONS AND RELEVANCE

A breakfast in the classroom initiative increased participation in the School Breakfast Program and did not affect the combined incidence of overweight and obesity. However, the initiative had an unintended consequence of increasing incident and prevalent obesity. Further research is needed to identify approaches to increase participation in the School Breakfast Program that do not increase obesity among students.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT01924130.

Not so implausible: impact of longitudinal assessment of implausible anthropometric measures on obesity prevalence and weight change in children and adolescents

PURPOSE

Implausible anthropometric measures are typically identified using population outlier definitions, conflating implausible and extreme measures. We determined the impact of a longitudinal outlier approach on prevalence of body mass index (BMI) categories and mean change in anthropometric measures in pediatric electronic health record data.

METHODS

We examined 996,131 observations from 147,375 children (10-18 years) in the ADVANCE Clinical Data Research Network, a national network of community health centers. Sex-stratified, mixed effects, linear spline regression modeled weight, height, and BMI as a function of age. Longitudinal outliers were defined as observations with studentized residual greater than |6|; population outliers were defined by Centers for Disease Control-defined z-score thresholds.

RESULTS

At least 99.7% of anthropometric measures were not extreme by longitudinal or population definitions (agreement ≥ 0.995). BMI category prevalence after excluding longitudinal or population outliers differed by less than 0.1%. Among children greater than 85th percentile at baseline, annual mean changes in anthropometric measures were larger in data that excluded longitudinal (girls: 1.24 inches, 12.39 pounds, 1.53 kg/m²; boys: 2.34, 14.08, 1.07) versus population outliers (girls: 0.61 inches, 8.22 pounds, 0.75 kg/m²; boys: 1.53, 11.61, 0.48).

CONCLUSIONS

Longitudinal outlier methods may reduce underestimation of anthropometric change in children with elevated baseline values.

Title: efficacy of a food parenting intervention for mothers with low income to reduce preschooler's solid fat and added sugar intakes: a randomized controlled trial

Background

Few interventions have shown efficacy to influence key energy balance behaviors during the preschool years.

Objective

A randomized controlled trial (RCT) was used to evaluate the efficacy of Food, Fun, and Families (FFF), a 12 week authoritative food parenting intervention for mothers with low-income levels, to reduce preschool-aged children’s intake of calories from solid fat and added sugar (SoFAS).

Methods

Mothers were randomly assigned to receive FFF (n = 59) or to a delayed treatment control (n = 60). The primary outcome was children’s daily energy intake from SoFAS at the end of the 12 week intervention, controlling for baseline levels, assessed by 24-h dietary recalls. Secondary outcomes included children’s daily energy intake, children’s BMI z-scores, and meal observations of maternal food parenting practices targeted in FFF (e.g. providing guided choices).

Results

Participating mothers were predominantly African American (91%), with 39% educated beyond high school and 66% unemployed. Baseline demographics and child SoFAS intakes did not differ by group. Lost to follow-up was 13% and did not differ between groups. At post-intervention, FFF children consumed ~ 94 kcal or 23% less daily energy from SoFAS than children in the control group, adjusting for baseline levels (307.8 (95%CI = 274.1, 341.5) kcal vs. 401.9 (95%CI = 369.8, 433.9) kcal, FFF vs. control; p < 0.001). FFF mothers also displayed a greater number of authoritative parenting practices when observed post-intervention with their child at a buffet-style meal (Wilcoxon z = − 2.54, p = 0.012). Neither child total daily energy intake nor BMI z-scores differed between groups post-intervention.

Conclusions

Findings demonstrate the initial efficacy of an authoritative food parenting intervention for families with low-income to reduce SoFAS intake in early childhood. Additional research is needed to evaluate longer-term effects on diet and growth.

Trial registration

Retrospectively registered at ClinicalTrials.gov: #NCT03646201.

Tracking and Variability in Childhood Levels of BMI: The Bogalusa Heart Study

OBJECTIVE

Although the tracking of BMI levels from childhood to adulthood has been examined, there is little information on the within-person variability of BMI.

METHODS

Longitudinal data from 11,591 schoolchildren, 3,096 of whom were reexamined as adults, were used to explore the tracking and variability of BMI levels. This article focuses on changes in age-adjusted levels of BMI.

RESULTS

There was strong tracking of BMI levels. The correlation of adjusted BMI levels was r = 0.88, and 78% of children with severe obesity at one examination had severe obesity at the next examination (mean interval, 2.7 years). Further, an increase in adjusted BMI from +5 kg/m² (above the median) to + 10 increased the risk for adult BMI ≥ 40 by 2.7-fold. However, BMI levels among children and adolescents were variable. Over a 9- to 15-month interval, the SD of adjusted BMI change was 0.9 kg/m² , and 0.7% of children had an absolute change ≥ 3.5. This variability was associated with the interval between examinations and with the initial BMI.

CONCLUSIONS

Despite the high degree of tracking of BMI, annual changes of 3.5 kg/m² or more are plausible. Knowledge of this variability is important when following a child over time.

Identifying and categorizing spurious weight data in electronic medical records

Background

Spurious weights compromise the validity of summary measures, such as averages and trends. Even rare errors in weight records can undermine the utility of electronic medical record (EMR) data.

Objective

We sought to estimate the prevalence of spurious weight values in a large EMR, to ascertain the likely causes, and to develop and test straightforward algorithms for identifying spurious weight data.

Design

Using EMR data from 10,000 randomly selected patients aged ≥65 y in the VA system, we examined the percentage of weight change across various time intervals, from 1 to 3000 d. We examined descriptive results and developed 3 algorithms to categorize degree of weight change over time. On the basis of distributions, we identified cases that were most likely spurious. We manually reviewed these and categorized the type of error.

Results

The data followed the expected distributions. The algorithms reliably identified spurious weight. Approximately 0.8% of all weights in the record appeared to be spurious and ∼1 in 5 patient charts included ≥1 spurious weight value. The most common type of error involved the misentry of a single digit (e.g., 148 for 178).

Conclusions

Spurious weights are common in EMRs. Straightforward algorithms can identify and remove them, and thus enhance the reliability of EMR data.

Automated identification of implausible values in growth data from pediatric electronic health records

Objective

Large electronic health record (EHR) datasets are increasingly used to facilitate research on growth, but measurement and recording errors can lead to biased results. We developed and tested an automated method for identifying implausible values in pediatric EHR growth data.

Materials and Methods

Using deidentified data from 46 primary care sites, we developed an algorithm to identify weight and height values that should be excluded from analysis, including implausible values and values that were recorded repeatedly without remeasurement. The foundation of the algorithm is a comparison of each measurement, expressed as a standard deviation score, with a weighted moving average of a child's other measurements. We evaluated the performance of the algorithm by (1) comparing its results with the judgment of physician reviewers for a stratified random selection of 400 measurements and (2) evaluating its accuracy in a dataset with simulated errors.

Results

Of 2 000 595 growth measurements from 280 610 patients 1 to 21 years old, 3.8% of weight and 4.5% of height values were identified as implausible or excluded for other reasons. The proportion excluded varied widely by primary care site. The automated method had a sensitivity of 97% (95% confidence interval [CI], 94-99%) and a specificity of 90% (95% CI, 85-94%) for identifying implausible values compared to physician judgment, and identified 95% (weight) and 98% (height) of simulated errors.

Discussion and Conclusion

This automated, flexible, and validated method for preparing large datasets will facilitate the use of pediatric EHR growth datasets for research.

New approach for the identification of implausible values and outliers in longitudinal childhood anthropometric data

PURPOSE

We aimed to demonstrate the use of jackknife residuals to take advantage of the longitudinal nature of available growth data in assessing potential biologically implausible values and outliers.

METHODS

Artificial errors were induced in 5% of length, weight, and head circumference measurements, measured on 1211 participants from the Maternal Vitamin D for Infant Growth (MDIG) trial from birth to 24 months of age. Each child's sex- and age-standardized z-score or raw measurements were regressed as a function of age in child-specific models. Each error responsible for a biologically implausible decrease between a consecutive pair of measurements was identified based on the higher of the two absolute values of jackknife residuals in each pair. In further analyses, outliers were identified as those values beyond fixed cutoffs of the jackknife residuals (e.g., greater than +5 or less than -5 in primary analyses). Kappa, sensitivity, and specificity were calculated over 1000 simulations to assess the ability of the jackknife residual method to detect induced errors and to compare these methods with the use of conditional growth percentiles and conventional cross-sectional methods.

RESULTS

Among the induced errors that resulted in a biologically implausible decrease in measurement between two consecutive values, the jackknife residual method identified the correct value in 84.3%-91.5% of these instances when applied to the sex- and age-standardized z-scores, with kappa values ranging from 0.685 to 0.795. Sensitivity and specificity of the jackknife method were higher than those of the conditional growth percentile method, but specificity was lower than for conventional cross-sectional methods.

CONCLUSIONS

Using jackknife residuals provides a simple method to identify biologically implausible values and outliers in longitudinal child growth data sets in which each child contributes at least 4 serial measurements.

Completeness and accuracy of anthropometric measurements in electronic medical records for children attending primary care

Background

Electronic medical records (EMRs) from primary care may be a feasible source of height and weight data. However, the use of EMRs in research has been impeded by lack of standardisation of EMRs systems, data access and concerns about the quality of the data.

Objectives

The study objectives were to determine the data completeness and accuracy of child heights and weights collected in primary care EMRs, and to identify factors associated with these data quality attributes.

Methods

A cross-sectional study examining height and weight data for children <19 years from EMRs through the Electronic Medical Record Administrative data Linked Database (EMRALD), a network of family practices across the province of Ontario. Body mass index z-scores were calculated using the World Health Organization Growth Standards and Reference.

Results

A total of 54,964 children were identified from EMRALD. Overall, 93% had at least one complete set of growth measurements to calculate a body mass index (BMI) z-score. 66.2% of all primary care visits had complete BMI z-score data. After stratifying by visit type 89.9% of well-child visits and 33.9% of sick visits had complete BMI z-score data; incomplete BMI z-score was mainly due to missing height measurements. Only 2.7% of BMI z-score data were excluded due to implausible values.

Conclusions

Data completeness at well-child visits and overall data accuracy were greater than 90%. EMRs may be a valid source of data to provide estimates of obesity in children who attend primary care.

Eating behavior traits associated with demographic variables and implications for obesity outcomes in early childhood

Despite ongoing investigation of children's eating behaviors to better understand the etiology of childhood obesity, few studies have aimed to determine differences in eating behavior based on psycho-social variables reflective of 'stressful' life circumstance. Cross-sectional data collected from parents of 977 Australian children (2.0-5.0 years) in an online survey was used to determine associations between parent-reported Children's Eating Behavior Questionnaire [CEBQ] sub-scales, child BMI z-scores and psycho-social variables. When examined individually, all CEBQ sub-scales, except Slowness in Eating, were associated with BMI z-score (Food Responsiveness B = 0.226, p = 0.003, Enjoyment of Food B = 0.169, p = 0.035, Food Fussiness B = -0.139, p = 0.024, Satiety Responsiveness B = -0.318, p = 0.001). On examining CEBQ sub-scales along with psycho-social demographic variables, only Food Responsiveness and Satiety Responsiveness were retained, along with being a boy, child age, and parent BMI. Food Responsiveness was positively associated with parental stress and child age and negatively with parent BMI, while Enjoyment of Food was positively associated with child sleep duration, single parent status, and negatively with breastfeeding less than 6 months and parental depression. Satiety Responsiveness was positively associated with parent BMI and child age, and negatively with child sleep duration, while Food Fussiness was positively associated with child age and breastfeeding less than 6 months, and negatively with sleep duration, parental depression and single parent status. Attention to eating behaviors and associated psycho-social variables may provide opportunity for targeted obesity prevention initiatives.