Behavioural patterns only predict concurrent BMI status and not BMI trajectories in a sample of youth in Ontario, Canada

A novel systematic pipeline for increased predictability and explainability of growth patterns in children using trajectory features

OBJECTIVE

Longitudinal patterns of growth in early childhood are associated with health conditions throughout life. Knowledge of such patterns and the ability to predict them can lead to better prevention and improved health promotion in adulthood. However, growth analyses are characterized by significant variability, and pattern detection is affected by the method applied. Moreover, pattern labelling is typically performed based on ad hoc methods, such as visualizations or clinical experience. Here, we propose a novel pipeline using features extracted from growth trajectories using mathematical, statistical and machine-learning approaches to predict growth patterns and label them in a systematic and unequivocal manner.

METHODS

We extracted mathematical and clinical features from 9577 children growth trajectories embedded with machine-learning predictions of the growth patterns. We experimented with two sets of features (CAnonical Time-series Characteristics and trajectory features specific to growth), developmental periods and six machine-learning classifiers. Clinical experts provided labels for the detected patterns and decision rules were created to associate the features with the labelled patterns. The predictive capacity of the extracted features was validated on two heterogenous populations (The Applied Research Group for Kids and the 2004 Pelotas Birth Cohort, based in Canada and Brazil, respectively).

RESULTS

Features predictive ability measured by accuracy and F1 score was ≥ 80% and ≥ 0.76 respectively in both cohorts. A small number of features (n = 74) was sufficient to distinguish between growth patterns in both cohorts. Slope, intercept of the trajectory, age at peak value, start value and change of the growth measure were among the top identified features.

CONCLUSION

Growth features can be reliably used as predictors of growth patterns and provide an unbiased understanding of growth patterns. They can be used as tool to reduce the effort to repeat analysis and variability concerning anthropometric measures, time points and analytical methods, in the context of the same or similar populations.

Clustering of Physical Activity, Diet and Sedentary Behavior among Youth from Low-, Middle-, and High-Income Countries: A Scoping Review

Background: The interaction between physical activity (PA), diet, and sedentary behavior (SB) plays an important role on health-related outcomes. This scoping review (Prospero CRD42018094826) aims to identify and appraise clusters of PA, diet, and SB among youth (0–19 years) according to country income. Methods: Five databases were searched. Fifty-seven articles met the inclusion criteria. Results: Fifty-five cluster types were identified, with greater variety in high-income than lower income countries. The most prevalent profiles were “High SB and consumption of sugar, salt, and beverages (SSB)” (n = 17) and “High PA” (n = 13–5), both of which presented in all income countries. The healthiest profile, “High PA and fruit and vegetables (F&V); Low SB and SSB” (n = 12), was present in upper-middle and high-income countries, while the unhealthiest “Low PA and F&V; High SB and SSB” (n = 6) was present only in high-income countries. Conclusions: High SB and unhealthy diet (SSB) were more prevalent in clusters, mainly in high-income countries. The results support the need for multi-component actions targeting more than one behavior at the same time.

Gender differences in the longitudinal association between multilevel latent classes of chronic disease risk behaviours and body mass index in adolescents

INTRODUCTION

Few studies have assessed the relationship between chronic disease risk behaviours and body mass index (BMI) in a longitudinal, sex/gender-specific context. This study used gender-specific analyses to assess the extent to which chronic disease risk behaviour latent classes are associated with BMI and weight status at follow-up.

METHODS

Longitudinal data from 4510 students in Grades 9 to 12, tracked from 2013- 2015, who participated in the COMPASS study were used to assess gender differences in the lagged association between previously determined latent classes (of physical activity and substance use) with BMI using multilevel mixed-effects models. Our multilevel regression models assessed the association between two latent classes, active experimenters and inactive non-using youth, with BMI when stratified by gender.

RESULTS

Male inactive non-substance-using youth were associated with a 0.29 higher continuous BMI (95% CI: 0.057, 0.53) and odds of overweight/obesity increased by 72% (OR = 1.72, 95% CI: 1.2, 2.4) for binary BMI at follow-up relative to active youth who experiment with substance use. No significant associations were detected in females.

CONCLUSION

Over time, physical activity has a protective role on BMI in male youth. Both substance use and physical inactivity should be addressed in obesity prevention efforts. Gender stratification in analyses is also important since females and males have different contributing factors to increases in BMI.

The global burden of disease attributable to high body mass index in 195 countries and territories, 1990-2017: An analysis of the Global Burden of Disease Study

BACKGROUND

Obesity represents an urgent problem that needs to be properly addressed, especially among children. Public and global health policy- and decision-makers need timely, reliable quantitative information to develop effective interventions aimed at counteracting the burden generated by high body mass index (BMI). Few studies have assessed the high-BMI-related burden on a global scale.

METHODS AND FINDINGS

Following the methodology framework and analytical strategies used in the Global Burden of Disease Study (GBD) 2017, the global deaths and disability-adjusted life years (DALYs) attributable to high BMI were analyzed by age, sex, year, and geographical location and by Socio-demographic Index (SDI). All causes of death and DALYs estimated in GBD 2017 were organized into 4 hierarchical levels: level 1 contained 3 broad cause groupings, level 2 included more specific categories within the level 1 groupings, level 3 comprised more detailed causes within the level 2 categories, and level 4 included sub-causes of some level 3 causes. From 1990 to 2017, the global deaths and DALYs attributable to high BMI have more than doubled for both females and males. However, during the study period, the age-standardized rate of high-BMI-related deaths remained stable for females and only increased by 14.5% for males, and the age-standardized rate of high-BMI-related DALYs only increased by 12.7% for females and 26.8% for males. In 2017, the 6 leading GBD level 3 causes of high-BMI-related DALYs were ischemic heart disease, stroke, diabetes mellitus, chronic kidney disease, hypertensive heart disease, and low back pain. For most GBD level 3 causes of high-BMI-related DALYs, high-income North America had the highest attributable proportions of age-standardized DALYs due to high BMI among the 21 GBD regions in both sexes, whereas the lowest attributable proportions were observed in high-income Asia Pacific for females and in eastern sub-Saharan Africa for males. The association between SDI and high-BMI-related DALYs suggested that the lowest age-standardized DALY rates were found in countries in the low-SDI quintile and high-SDI quintile in 2017, and from 1990 to 2017, the age-standardized DALY rates tended to increase in regions with the lowest SDI, but declined in regions with the highest SDI, with the exception of high-income North America. The study's main limitations included the use of information collected from some self-reported data, the employment of cutoff values that may not be adequate for all populations and groups at risk, and the use of a metric that cannot distinguish between lean and fat mass.

CONCLUSIONS

In this study, we observed that the number of global deaths and DALYs attributable to high BMI has substantially increased between 1990 and 2017. Successful population-wide initiatives targeting high BMI may mitigate the burden of a wide range of diseases. Given the large variations in high-BMI-related burden of disease by SDI, future strategies to prevent and reduce the burden should be developed and implemented based on country-specific development status.

Longitudinal changes in body composition and waist circumference by self-reported levels of physical activity in leisure among adolescents: the Tromsø study, Fit Futures

BACKGROUND

It is not clear how physical activity affects body composition in adolescents. Physical activity levels are often reduced during this period, and the relative proportion of body fat mass and lean mass undergo natural changes in growing adolescents. We aimed to examine whether self-reported physical activity in leisure time at baseline or change in activity during follow-up affect changes in four measures of body composition; body mass index (kg/m2), waist circumference, fat mass index (fat mass in kg/m2) and lean mass index (lean mass in kg/m2).

METHODS

We used data from the Tromsø Study Fit Futures, which invited all first year students in upper secondary high school in two municipalities in northern Norway in 2010-2011. They were reexamined in 2012-2013. Longitudinal data was available for 292 boys and 354 girls. We used multiple linear regression analyses to assess whether self-reported level of physical activity in leisure time at baseline predicted changes in body composition, and analysis of covariance to assess the effects of change in level of activity during follow-up on change in body composition. All analyses were performed sex-specific, and a p-value of < 0.05 was considered statistically significant.

RESULTS

There were no associations between self-reported leisure time physical activity in the first year of upper secondary high school and changes in any of the considered measure of body composition after 2 years of follow up, with the exception of waist circumference in boys (p = 0.05). In boys, change in fat mass index differed significantly between groups of activity change (p < 0.01), with boys adopting activity or remaining physically active having less increase in fat mass index than the consistently inactive. In girls, change in lean mass index differed significantly between groups of activity change (p = 0.04), with girls adopting physical activity having the highest increase.

CONCLUSIONS

Self-reported leisure time physical activity does not predict changes in body composition in adolescents after 2 years of follow up. Change in the level of physical activity is associated with change in fat mass index in boys and lean mass index in girls.

Activity-related behavior typologies in youth: a systematic review

BACKGROUND

Clusters of adolescents differentiated by patterns of physical activity and sedentary behavior (activity-related typologies) are common. Understanding both the characteristics of adolescents and modifiable correlates of these typologies, can help to develop interventions for those most at risk. This systematic review aimed to synthesize the socio-demographic characteristics and modifiable correlates of activity-related behavioral typologies among adolescents.

METHODS

A systematic search of seven electronic databases was conducted to identify quantitative studies using person-oriented statistical approaches to identify activity-related behavioral typologies among 12-18 year-olds. This systematic review was registered in Prospero (registration number: CRD42016046879).

RESULTS

Thirty-six studies met the inclusion criteria and were classified according to three sub-themes based on behaviors included in the typologies (1. physical activity and sedentary behavior only; 2. physical activity, sedentary behavior and risk-related behaviors; 3. physical activity, sedentary behavior and diet). Studies were mostly cross-sectional and relied on self-report measures. Methods were considerably heterogeneous, however results revealed some consistency in typologies within specific groups. For example, typologies characterized by unhealthy behavior patterns (e.g., characterized by physical inactivity, high sedentary behavior and poor diet or high risk-related behaviors) comprised more older adolescents. With the exception of socio-demographics (age, sex, body mass index and socio-economic status), very few correlates have been studied to date (mostly school-related behavioral factors and intrapersonal influences), with evidence largely from typologies comprised of physical activity, sedentary behavior and diet.

CONCLUSIONS

More research is needed to assess a range of modifiable correlates associated with activity-related behavior typologies among adolescents. This will allow for more targeted interventions, to achieve long-lasting, positive behavior change in adolescent populations.