Longitudinal drop-out and weighting against its bias

Developing non-response weights to account for attrition-related bias in a longitudinal pregnancy cohort

BACKGROUND

Prospective cohorts may be vulnerable to bias due to attrition. Inverse probability weights have been proposed as a method to help mitigate this bias. The current study used the "All Our Families" longitudinal pregnancy cohort of 3351 maternal-infant pairs and aimed to develop inverse probability weights using logistic regression models to predict study continuation versus drop-out from baseline to the three-year data collection wave.

METHODS

Two methods of variable selection took place. One method was a knowledge-based a priori variable selection approach, while the second used Least Absolute Shrinkage and Selection Operator (LASSO). The ability of each model to predict continuing participation through discrimination and calibration for both approaches were evaluated by examining area under the receiver operating curve (AUROC) and calibration plots, respectively. Stabilized inverse probability weights were generated using predicted probabilities. Weight performance was assessed using standardized differences of baseline characteristics for those who continue in study and those that do not, with and without weights (unadjusted estimates).

RESULTS

The a priori and LASSO variable selection method prediction models had good and fair discrimination with AUROC of 0.69 (95% Confidence Interval [CI]: 0.67-0.71) and 0.73 (95% CI: 0.71-0.75), respectively. Calibration plots and non-significant Hosmer-Lemeshow Goodness of Fit Tests indicated that both the a priori (p = 0.329) and LASSO model (p = 0.242) were well-calibrated. Unweighted results indicated large (> 10%) standardized differences in 15 demographic variables (range: 11 - 29%), when comparing those who continued in the study with those that did not. Weights derived from the a priori and LASSO models reduced standardized differences relative to unadjusted estimates, with the largest differences of 13% and 5%, respectively. Additionally, when applying the same LASSO variable selection method to develop weights in future data collection waves, standardized differences remained below 10% for each demographic variable.

CONCLUSION

The LASSO variable selection approach produced robust weights that addressed non-response bias more than the knowledge-driven approach. These weights can be applied to analyses across multiple longitudinal waves of data collection to reduce bias.

Identification of high likelihood of dementia in population-based surveys using unsupervised clustering: a longitudinal analysis

BACKGROUND

Dementia is defined as a cognitive decline that affects functional status. Longitudinal ageing surveys often lack a clinical diagnosis of dementia though measure cognition and daily function over time. We used unsupervised machine learning and longitudinal data to identify transition to probable dementia.

METHODS

Multiple Factor Analysis was applied to longitudinal function and cognitive data of 15,278 baseline participants (aged 50 years and more) from the Survey of Health, Ageing, and Retirement in Europe (SHARE) (waves 1, 2 and 4-7, between 2004 and 2017). Hierarchical Clustering on Principal Components discriminated three clusters at each wave. We estimated probable or "Likely Dementia" prevalence by sex and age, and assessed whether dementia risk factors increased the risk of being assigned probable dementia status using multistate models. Next, we compared the "Likely Dementia" cluster with self-reported dementia status and replicated our findings in the English Longitudinal Study of Ageing (ELSA) cohort (waves 1-9, between 2002 and 2019, 7840 participants at baseline).

RESULTS

Our algorithm identified a higher number of probable dementia cases compared with self-reported cases and showed good discriminative power across all waves (AUC ranged from 0.754 [0.722-0.787] to 0.830 [0.800-0.861]). "Likely Dementia" status was more prevalent in older people, displayed a 2:1 female/male ratio, and was associated with nine factors that increased risk of transition to dementia: low education, hearing loss, hypertension, drinking, smoking, depression, social isolation, physical inactivity, diabetes, and obesity. Results were replicated in ELSA cohort with good accuracy.

CONCLUSIONS

Machine learning clustering can be used to study dementia determinants and outcomes in longitudinal population ageing surveys in which dementia clinical diagnosis is lacking.

Short-Term Quality of Life after Autologous Compared with Alloplastic Breast Reconstruction: A Prospective Study

BACKGROUND

The aim of this prospective multicenter study was to evaluate whether autologous breast reconstruction (BR) leads to lower short-term quality of life (QoL) compared with alloplastic BR, due to the more physically demanding surgery and increased risk of severe complications of autologous BR.

METHODS

Changes in QoL after BR were measured in this prospective multicenter study using the BREAST-Q questionnaire, which was administered preoperatively and at 6 weeks and 6 months postoperatively. Characteristics and complications, classified according to Clavien-Dindo, were compared between alloplastic and autologous groups. Profile plots and generalized linear regression models were constructed to analyze the BREAST-Q subscales over time for both BR groups.

RESULTS

Preoperatively, women undergoing autologous BR scored lower on all BREAST-Q scales compared with women undergoing alloplastic BR, regardless of whether they underwent immediate or delayed BR. Women undergoing autologous BR scored higher at 6 weeks and 6 months postoperatively on the Satisfaction with Breasts ( P = 0.001), Psychosocial Well-Being ( P = 0.024), and Sexual Well-Being ( P = 0.007) subscales. Postoperative Physical Well-Being: Chest score was similar between the groups ( P = 0.533). Clavien-Dindo grade III or higher complications occurred more often among women in the autologous group (27% versus 12%, P = 0.042). Complications were not associated with worse BREAST-Q scores on any of the subscales.

CONCLUSIONS

In contrast to the authors' expectations, and despite the higher incidence of severe complications and lower preoperative breast satisfaction and QoL scores, women undergoing autologous BR had higher levels of breast satisfaction and psychosocial and sexual well-being, both at 6 weeks and 6 months after BR, compared with women undergoing alloplastic BR.

CLINICAL QUESTION/LEVEL OF EVIDENCE

Therapeutic, II.

Estimating the causal effects of work-related and non-work-related stressors on perceived stress level: A fixed effects approach using population-based panel data

OBJECTIVES

Prolonged or excessive stress can have a negative impact on health and well-being, and stress therefore constitutes a major public health issue. A central question is what are the main sources of stress in contemporary societies? This study examines the effects of work-related and non-work-related stressors and perceived social support on perceived stress within a causal framework.

METHODS

Panel data were drawn from two waves (2013 and 2017) of the population-based health survey "How are you?" conducted in the Central Denmark Region. The analytical sample comprised 9,194 subjects who had responded to both surveys. Work-related and non-work-related stressors included major life events, chronic stressors, daily hassles and lack of social support. Perceived stress was measured with the 10-item Perceived Stress Scale (PSS). Data were analysed using fixed effects regression in a fully balanced design.

RESULTS

The largest effects on PSS were seen in own disease, work situation and lack of social support. Other stressors affecting the perceived stress level were financial circumstances, relationship with partner, relationship with family and friends, and disease among close relatives. Most variables had a symmetrical effect on PSS.

CONCLUSIONS

The results point to the need for comprehensive policies to promote mental health that span life domains and include both the individual and the group as well as organizational and societal levels. The study indicates that there are multiple potential entry points for stress prevention and stress management. However, it also shows that disease, work situation and social support weigh heavily in the overall picture. This points to the healthcare system and workplace as key institutional venues for action.

Childhood Abuse, Intimate Partner Violence in Young Adulthood, and Welfare Receipt by Midlife

OBJECTIVE

To investigate prospective associations between type of child abuse (physical, sexual, both), timing (childhood, young adulthood, both), and welfare receipt into middle-age.

METHODS

Database linkage study using the Quebec Longitudinal Study of Kindergarten Children cohort born in 1980 and government administrative databases (N = 3020). We assessed parental tax returns, family and personal background characteristics (1982-1987). At age 22 years, participants answered retrospective questionnaires on experienced childhood abuse (physical, sexual abuse < age 18 years) and intimate partner violence (IPV) (ages 18-22). Main outcome was years on social assistance, on the basis of participant tax returns (ages 23-37 years). Analysis included weights for population representativeness.

RESULTS

Of 1690 participants (54.4% females) with available data, 22.4% reported childhood abuse only, 14.5% IPV only, and 18.5% both. Prevalence of childhood physical, sexual, and both was 20.4%, 12.2%, and 8.3%, respectively. Adjusting for socioeconomic background and individual characteristics, we found that childhood physical abuse alone and physical or sexual abuse combined were associated with a two-fold risk of welfare receipt, as compared to never-abused (adjusted incidence risk ratio 2.43, 95% confidence interval [CI], 1.65-3.58; and adjusted incidence risk ratio 2.04, 95% CI, 1.29-3.23, respectively). Repeated abuse (childhood abuse combined with adult IPV) had a three-fold risk (adjusted incidence ratio 3.59, 95% CI, 2.39-5.37).

CONCLUSIONS

Abuse across several developmental periods (childhood and young adulthood) is associated with increased risks of long-term welfare receipt, independently of socioeconomic background. Results indicate a dose-response association. Early prevention and targeted identification are crucial to preventing economic adversity that may potentially lead to intergenerational poverty.

Identification of High Likelihood of Dementia in Population-Based Surveys using Unsupervised Clustering: a Longitudinal Analysis

Background

Dementia is defined by cognitive decline that affects functional status. Longitudinal ageing surveys often lack a clinical diagnosis of dementia though measure cognitive and function over time. We used unsupervised machine learning and longitudinal data to identify transition to probable dementia.

Methods

Multiple Factor Analysis was applied to longitudinal function and cognitive data of 15,278 baseline participants (aged 50 years and more) from the Survey of Health, Ageing, and Retirement in Europe (SHARE) (waves 1, 2 and 4-7, between 2004 and 2017). Hierarchical Clustering on Principal Components discriminated three clusters at each wave. We estimated probable or "Likely Dementia" prevalence by sex and age, and assessed whether dementia risk factors increased the risk of being assigned probable dementia status using multistate models. Next, we compared the "Likely Dementia" cluster with self-reported dementia status and replicated our findings in the English Longitudinal Study of Ageing (ELSA) cohort (waves 1-9, between 2002 and 2019, 7,840 participants at baseline).

Findings

Our algorithm identified a higher number of probable dementia cases compared with self-reported cases and showed good discriminative power across all waves (AUC ranged from 0.754 [0.722-0.787] to 0.830 [0.800-0.861]). "Likely Dementia" status was more prevalent in older people, displayed a 2:1 female/male ratio and was associated with nine factors that increased risk of transition to dementia: low education, hearing loss, hypertension, drinking, smoking, depression, social isolation, physical inactivity, diabetes, and obesity. Results were replicated in ELSA cohort with good accuracy.

Interpretation

Machine learning clustering can be used to study dementia determinants and outcomes in longitudinal population ageing surveys in which dementia clinical diagnosis is lacking.

Why weight? Analytic approaches for large-scale population neuroscience data

Population-based neuroimaging studies that feature complex sampling designs enable researchers to generalize their results more widely. However, several theoretical and analytical questions pose challenges to researchers interested in these data. The following is a resource for researchers interested in using population-based neuroimaging data. We provide an overview of sampling designs and describe the differences between traditional model-based analyses and survey-oriented design-based analyses. To elucidate key concepts, we leverage data from the Adolescent Brain Cognitive Development℠ Study (ABCD Study®), a population-based sample of 11,878 9-10-year-olds in the United States. Analyses revealed modest sociodemographic discrepancies between the target population of 9-10-year-olds in the U.S. and both the recruited ABCD sample and the analytic sample with usable structural and functional imaging data. In evaluating the associations between socioeconomic resources (i.e., constructs that are tightly linked to recruitment biases) and several metrics of brain development, we show that model-based approaches over-estimated the associations of household income and under-estimated the associations of caregiver education with total cortical volume and surface area. Comparable results were found in models predicting neural function during two fMRI task paradigms. We conclude with recommendations for ABCD Study® users and users of population-based neuroimaging cohorts more broadly.

Uptake and impact of the English National Health Service digital diabetes prevention programme: observational study

INTRODUCTION

'Healthier You', the National Health Service (NHS) diabetes prevention programme (DPP) offers adults in England at high risk of type 2 diabetes (T2DM) an evidence-based behavioral intervention to prevent or delay T2DM onset. This study assesses the impact of a pilot digital stream of the DPP (DDPP) on glycated hemoglobin (HbA1c) and weight.

RESEARCH DESIGN AND METHODS

A service evaluation employing prospectively collected data in a prospective cohort design in nine NHS local pilot areas across England. Participants were adults with non-diabetic hyperglycemia (NDH) (HbA1c 42-47 mmol/mol or fasting plasma glucose 5.5-6.9 mmol/L) in the 12 months prior to referral. The DDPP comprised five digital health interventions (DHI). Joint primary outcomes were changes in HbA1c and weight between baseline and 12 months. HbA1c and weight readings were recorded at referral (baseline) by general practices, and then at 12-month postregistration. Demographic data and service variables were collected from the DHI providers.

RESULTS

3623 participants with NDH registered for the DDPP and of these, 2734 (75%) were eligible for inclusion in the analyses. Final (12-month) follow-up data for HbA1c were available for 1799 (50%) and for weight 1817 (50%) of registered participants. Mean change at 12 months was -3.1 (-3.4 to -2.8) kg, p<0.001 for weight and -1.6 (-1.8 to -1.4) mmol/mol, p<0.001 for HbA1c. Access to peer support and a website and telephone service was associated with significantly greater reductions in HbA1c and weight.

CONCLUSIONS

Participation in the DDPP was associated with clinically significant reductions in weight and HbA1c. Digital diabetes prevention can be an effective and wide-reaching component of a population-based approach to addressing type 2 diabetes prevention.

A Joint Model for Unbalanced Nested Repeated Measures with Informative Drop-Out Applied to Ambulatory Blood Pressure Monitoring Data

This study proposes a Bayesian joint model with extended random effects structure that incorporates nested repeated measures and provides simultaneous inference on treatment effects over time and drop-out patterns. The proposed model includes flexible splines to characterize the circadian variation inherent in blood pressure sequences, and we assess the effectiveness of an intervention to resolve pediatric obstructive sleep apnea. We demonstrate that the proposed model and its conventional two-stage counterpart provide similar estimates of nighttime blood pressure but estimates on the mean evolution of daytime blood pressure are discrepant. Our simulation studies tailored to the motivating data suggest reasonable estimation and coverage probabilities for both fixed and random effects. Computational challenges of model implementation are discussed.

Validation and generalizability of machine learning prediction models on attrition in longitudinal studies

Attrition in longitudinal studies is a major threat to the representativeness of the data and the generalizability of the findings. Typical approaches to address systematic nonresponse are either expensive and unsatisfactory (e.g., oversampling) or rely on the unrealistic assumption of data missing at random (e.g., multiple imputation). Thus, models that effectively predict who most likely drops out in subsequent occasions might offer the opportunity to take countermeasures (e.g., incentives). With the current study, we introduce a longitudinal model validation approach and examine whether attrition in two nationally representative longitudinal panel studies can be predicted accurately. We compare the performance of a basic logistic regression model with a more flexible, data-driven machine learning algorithm—gradient boosting machines. Our results show almost no difference in accuracies for both modeling approaches, which contradicts claims of similar studies on survey attrition. Prediction models could not be generalized across surveys and were less accurate when tested at a later survey wave. We discuss the implications of these findings for survey retention, the use of complex machine learning algorithms, and give some recommendations to deal with study attrition.

Reference centiles based on year-to-year changes for a longitudinal evaluation of motor performance in children and adolescents

Objectives

The aim was to evaluate the longitudinal course of motor skills development in children with a view to improve the understanding of intra-individual variance. Previous publications have been based on cross-sectional data or analyzed longitudinal studies in a cross-sectional manner.

Design

Longitudinal. Year-to-year change.

Methods

The present study is based on motor function data collected in the LIFE Child study (Germany). The participants (6 to 17 years) completed parts of the motor tests according to the standard of the German Motor Performance Test 6–18 (DMT). For a total of 1653 participants, 4616 motor tests with an annual interval in the period 2011 to 2019 were included in the evaluation.

Results

We were able to produce gender and age-specific change centiles for the test items standing long jump, jumping sideways, push-ups, stand and reach and balancing walking backwards. Each set of centiles revealed a range of variability in motor development in children and adolescents, with distinct progressive patterns in the different test items and different genders. The supplied tables offer an indication of expected year-to-year change for each test item depending on age and gender. Depending on the test item and the mean age, a deterioration in test results after a one-year interval was observed, despite cross-sectional centiles showing an upward trend.

Conclusion

We present a user-friendly tool as a way to assess individual dynamic changes in motor development of children and adolescents. In combination with the well-known cross-sectional centiles as baseline, this can be helpful for the scientific evaluation of motor skills tests and can also be used in school settings.

Physical fitness of children and youth with asthma in comparison to the reference population : Cross-sectional results of the population-based MoMo study in Germany

BACKGROUND

Physical fitness is an essential marker of health. The literature regarding the question of whether individuals with asthma have reduced physical fitness compared to their non-asthmatic peers is inconsistent and focuses on the cardiorespiratory endurance dimension. This study provides a comparison of different dimensions of physical fitness in individuals with and without asthma on the basis of the German population-based study "KiGGS" (German Health Interview and Examination Survey for Children and Adolescents) and its in-depth study "MoMo" (2009-2012: wave 1 and 2014-2017: wave 2).

METHODS

In total, 7731 individuals aged 6-30 years were included in this cross-sectional analysis at two measurement waves, including 353 individuals with and 7378 without asthma. The 12-month prevalence of physician-diagnosed asthma was assessed by interview. Physical fitness was measured by six test items of the MoMo test profile. "Cardiorespiratory endurance" was measured by an ergometric test, "muscular strength" by standing long jump, push-ups and sit-ups and "coordination" by jumping sideways and balancing backwards. Because of the broad age range of the sample, age- and sex-specific percentiles were used. Physical activity, age, gender and general state of health were assessed by questionnaire.

RESULTS

The individuals with asthma reported a poorer general state of health at both measurement waves. However, the results of the fitness tests indicated that they were as physically fit as their peers without asthma in relation to cardiorespiratory endurance and muscular strength. The mean percentiles were all within the same range. The results of the comparisons of coordination performance were inconsistent. At wave 1 they were within the same range, at wave 2 individuals with asthma showed a poorer coordination performance (p = 0.041; HL = 4.125, CI of HL 0.155-8.125).

CONCLUSIONS

To the best of our knowledge, this is the first study to compare the physical fitness of individuals with and without asthma by considering several dimensions of physical fitness. The study demonstrates that cardiorespiratory endurance and muscular strength are not reduced in individuals with asthma. The results of the comparisons at the two measurement waves were remarkably stable.

Parenting in poor health: Examining associations between parental health, prescription drug use, and child maltreatment

RATIONALE

Child maltreatment and problematic parenting are related to negative outcomes for children. Poor parental health could be a risk factor for problematic parenting through several mechanisms: 1) inadequate emotional regulation and coping; 2) impairment of parental capacity; and, 3) impairment of the parent-child relationship.

OBJECTIVE

This study examines relationships between self-rated parental health, prescription drug use, and a broad array of negative parenting outcomes.

METHODS

A sample of general population parents of children aged ten and younger was recruited from 30 mid-sized cities in California (n = 681). Weighted mixed-effects negative binomial and logistic regression models were used to examine associations between poor parental health, prescription drug use and child maltreatment (physical abuse, supervisory neglect, and physical neglect), and problematic parenting (psychological aggression and corporal punishment).

RESULTS

Parents in poor health used physical abuse, corporal punishment, and psychological aggression more frequently and had higher odds of supervisory neglect. Parents who were taking more prescription medications had higher odds of physical neglect. Exploratory analyses suggested that prescriptions for certain medical conditions both increased and decreased the risk of problematic parenting.

CONCLUSIONS

Poor health and prescription drug use are not uncommon and present largely under-recognized risk factors for a spectrum of adverse parenting outcomes. Our study provides additional evidence that parents in poor health are at heightened risk of negative parenting, and need targeted intervention supports to support family well-being.

Choosing the right time granularity for analysis of digital biomarker trajectories

INTRODUCTION

The use of digital biomarker data in dementia research provides the opportunity for frequent cognitive and functional assessments that was not previously available using conventional approaches. Assessing high-frequency digital biomarker data can potentially increase the opportunities for early detection of cognitive and functional decline because of improved precision of person-specific trajectories. However, we often face a decision to condense time-stamped data into a coarser time granularity, defined as the frequency at which measurements are observed or summarized, for statistical analyses. It is important to find a balance between ease of analysis by condensing data and the integrity of the data, which is reflected in a chosen time granularity.

METHODS

In this paper, we discuss factors that need to be considered when faced with a time granularity decision. These factors include follow-up time, variables of interest, pattern detection, and signal-to-noise ratio.

RESULTS

We applied our procedure to real-world data which include longitudinal in-home monitored walking speed. The example shed lights on typical problems that data present and how we could use the above factors in exploratory analysis to choose an appropriate time granularity.

DISCUSSION

Further work is required to explore issues with missing data and computational efficiency.

Reweighting to address nonparticipation and missing data bias in a longitudinal electronic health record study

PURPOSE

We examined whether weighting techniques could account for longitudinal differences in disease activity by race/ethnicity between research participants and nonparticipants with rheumatoid arthritis (RA).

METHODS

We included 377 patients with RA from a public hospital in San Francisco, CA. We estimated the probability of not enrolling in a research study by constructing weights using inverse probability weighting. Disease activity over time by race/ethnicity was analyzed across the entire patient population and among research participants only using multivariable mixed-effects models.

RESULTS

There were no differences in RA disease activity scores between research participants and nonparticipants at baseline; however, longitudinal differences in disease activity between research participants and nonparticipants were found by race/ethnicity. Weighting research participants in accordance with sociodemographic and clinical characteristics of the nonparticipant population did not result in any meaningful changes in disease activity by race/ethnicity over time.

CONCLUSIONS

In our study of patients with RA, inverse probability weighting using select sociodemographic and clinical variables was not sufficient to account for longitudinal disease activity differences by race/ethnicity between research participants and nonparticipants.

Population-based trends in physical fitness of children and adolescents in Germany, 2003-2017

Physical fitness is an important health resource. From 2003 onwards, the MoMo study, a part of the German Health Interview and Examination Survey for Children and Adolescents (KiGGS), has investigated physical fitness on the basis of nationally representative data. The objective of this paper was to analyse trends in physical fitness of three cohorts at the ages of 4-17 years. Changes within 10 test items of the MoMo-test profile are reported. The mean values of physical fitness and the relevant influencing factors were directly compared across cohorts (in the total group and in age-and sex-subgroups). Statistical significance of differences was examined by calculating the confidence intervals (95% CI) for complex samples; effect sizes were determined using Cohen's d. The results indicated that physical fitness levels of German children and adolescents initially slightly increased between MoMo baseline (2003-2006) and wave 1 (2009-2012). This could partly be attributed to increased physical activity in sports clubs and in extracurricular sports at school. Then, between MoMo wave 1 and MoMo wave 2 (2014-2017) the physical fitness of children and adolescents in Germany remained the same, as did all influencing factors with the exception of extracurricular sport. From a public health perspective, programmes and measures that can increase the chances for all children and adolescents to strive for greater physical fitness should be of the highest priority.

The physical activity of children and adolescents in Germany 2003-2017: The MoMo-study

With digitalization and virtual entertainment being the megatrends of the 21st century, there is reasonable concern about the role of physical activity (PA) in the daily life of children and adolescents. To identify risk-groups with insufficient PA and to guide interventions, continuous and representative tracking of PA is crucial. In this paper, representative PA data of children and adolescents from the Motorik-Modul (MoMo) baseline study (2003-2006, N = 4,528) is compared to those of Wave 2 (2014-2017, N = 3,708). Participants aged 4-17 were drawn out of 167 sample points in Germany and the data was weighted to ensure representativeness for Germany. Organized (sports clubs and schools) and unorganized (unorganized sports and playing outside) PA was measured by questionnaire and stratified by sex, age, and socioeconomic status. Contrary to common expectation, overall PA remained stable among youths in the past ten years, however, there is an ongoing trend towards organized forms of PA at the expense of unorganized sports and playing outside. Besides different trends in settings, there is inequality in PA distribution among socioeconomic status and gender, unequally pronounced in different settings.

Social Support and Modelling in Relation to Physical Activity Participation and Outdoor Play in Preschool Children

Physical activity during early childhood is a prerequisite for healthy development in many cases. The aim of this study was to assess the relationships of social modelling and support from parents, peers, and siblings and domain-specific physical activity participation in a nationwide sample of preschool boys and girls from Germany. 519 preschool children aged 4-6 and one of their parents participated in the 'MoMo' Wave 1 Study between 2009 and 2012. Participants and their parents provided self-reported data on social support modelling, and domain-specific physical activity participation (physical activity in sports clubs, physical activity outside of sports clubs, and outdoor play). Parental, peer, and sibling support and modelling were related to domain-specific physical activity: Parental support was particularly relevant for physical activity in sports clubs, and peer support for outdoor play. Parental modelling was only related to physical activity outside of sports clubs: Maternal modelling was a positive correlate in girls and paternal modelling in boys, respectively. Sibling and peer modelling were especially relevant for physical activity in sports clubs. The results were heterogeneous regarding types and providers of support and modelling. Thus, different providers and types of support should be targeted in physical activity promotion programs for preschool children.

Representative body composition percentiles from bioelectrical impedance analyses among children and adolescents. The MoMo study

BACKGROUND & AIMS

With high prevalence of overweight and obesity in developed countries, the assessment of body composition in children and adolescents is of growing interest. The aim of this study was (i) to provide up-to-date population wide percentiles for meaningful diagnostic bioelectrical impedance outcome parameters and (ii) to specify reference values for these measures in a normal weight subpopulation.

METHODS

A representative nationwide sample of 4 to 24-year-old children and adolescents (n = 3593) was drawn out of 167 primary sample points across Germany from 2009 to 2012 (Motorik-Modul study wave 1). Fat free mass index (FFMI), fat mass index (FMI), body cell mass index (BCMI) and phase angle (PA) were derived from four electrodes 50 kHz bioelectrical impedance analyses. Sex-specific percentiles were calculated using the LMS-method for the total study population and a normal weight subsample (defined according to the international BMI-standard by Cole).

RESULTS

Percentile curves for body composition parameters are similar between boys and girls until puberty. Subsequently, girls show a higher FMI than boys, and boys increase their FFM, BCM, and PA time-shifted, in that order. Differences in FMI between the overall and the normal weight sample increase with age, showing an age-dependent prevalence for overweight and obesity among the representative German sample.

CONCLUSIONS

Results from the normal weight sample can be used as reference values to define a healthy body composition. Differences in body composition between the representative German sample and the normal weight subsample reveal a high prevalence of elevated fat mass in the population of children and adolescents living in Germany.