scholarly journals Does Physical Activity Predict Obesity—A Machine Learning and Statistical Method-Based Analysis

Author(s):  
Xiaolu Cheng ◽  
Shuo-yu Lin ◽  
Jin Liu ◽  
Shiyong Liu ◽  
Jun Zhang ◽  
...  

Background: Obesity prevalence has become one of the most prominent issues in global public health. Physical activity has been recognized as a key player in the obesity epidemic. Objectives: The objectives of this study are to (1) examine the relationship between physical activity and weight status and (2) assess the performance and predictive power of a set of popular machine learning and traditional statistical methods. Methods: National Health and Nutrition Examination Survey (NHANES, 2003 to 2006) data were used. A total of 7162 participants met our inclusion criteria (3682 males and 3480 females), with average age ranging from 48.6 (normal weight) to 52.1 years old (overweight). Eleven classifying algorithms—including logistic regression, naïve Bayes, Radial Basis Function (RBF), local k-nearest neighbors (k-NN), classification via regression (CVR), random subspace, decision table, multiobjective evolutionary fuzzy classifier, random tree, J48, and multilayer perceptron—were implemented and evaluated, and they were compared with traditional logistic regression model estimates. Results: With physical activity and basic demographic status, of all methods analyzed, the random subspace classifier algorithm achieved the highest overall accuracy and area under the receiver operating characteristic (ROC) curve (AUC). The duration of vigorous-intensity activity in one week and the duration of moderate-intensity activity in one week were important attributes. In general, most algorithms showed similar performance. Logistic regression was middle-ranking in terms of overall accuracy, sensitivity, specificity, and AUC among all methods. Conclusions: Physical activity was an important factor in predicting weight status, with gender, age, and race/ethnicity being less but still essential factors associated with weight outcomes. Tailored intervention policies and programs should target the differences rooted in these demographic factors to curb the increase in the prevalence of obesity and reduce disparities among sub-demographic populations.

Cholesterol ◽  
2012 ◽  
Vol 2012 ◽  
pp. 1-5 ◽  
Author(s):  
Fereshteh Baygi ◽  
Ahmad Reza Dorosty ◽  
Roya Kelishadi ◽  
Mostafa Qorbani ◽  
Hamid Asayesh ◽  
...  

Childhood obesity has become, a global public health problem, and epidemiological studies are important to identify its determinants in different populations. This study aimed to investigate factors associated with obesity in a representative sample of children in Neishabour, Iran. This study was conducted among 1500 randomly selected 6–12-year-old students from urban areas of Neishabour, northeast of Iran. Then, through a case-control study, 114 obese (BMI≥95th percentile of Iranian reference) children were selected as the case group and were compared with 102 controls (15th≤BMI<85th percentile). Factors suggested to be associated with weight status were investigated, for example, parental obesity, child physical activity levels, socio-economic status (SES), and so forth. The analysis was conducted using univariate and multivariate logistic regression (MLR) in SPSS version 16. In univariate logistic regression model, birth weight, birth order, family extension, TV watching, sleep duration, physical activity, parents’ job, parents’ education, parental obesity history, and SES were significantly associated with children’s obesity. After MLR analysis, physical activity and parental obesity history remained statistically significant in the model. Our findings showed that physical activity and parental obesity history are the most important determinants for childhood obesity in our population. This finding should be considered in implementation of preventive interventions.


Author(s):  
Keith Brazendale ◽  
◽  
Michael W. Beets ◽  
Bridget Armstrong ◽  
R. Glenn Weaver ◽  
...  

Abstract Purpose The Structured Days Hypothesis (SDH) posits that children’s behaviors associated with obesity – such as physical activity – are more favorable on days that contain more ‘structure’ (i.e., a pre-planned, segmented, and adult-supervised environment) such as school weekdays, compared to days with less structure, such as weekend days. The purpose of this study was to compare children’s moderate-to-vigorous physical activity (MVPA) levels on weekdays versus weekend days using a large, multi-country, accelerometer-measured physical activity dataset. Methods Data were received from the International Children’s Accelerometer Database (ICAD) July 2019. The ICAD inclusion criteria for a valid day of wear, only non-intervention data (e.g., baseline intervention data), children with at least 1 weekday and 1 weekend day, and ICAD studies with data collected exclusively during school months, were included for analyses. Mixed effects models accounting for the nested nature of the data (i.e., days within children) assessed MVPA minutes per day (min/day MVPA) differences between weekdays and weekend days by region/country, adjusted for age, sex, and total wear time. Separate meta-analytical models explored differences by age and country/region for sex and child weight-status. Results/findings Valid data from 15 studies representing 5794 children (61% female, 10.7 ± 2.1 yrs., 24% with overweight/obesity) and 35,263 days of valid accelerometer data from 5 distinct countries/regions were used. Boys and girls accumulated 12.6 min/day (95% CI: 9.0, 16.2) and 9.4 min/day (95% CI: 7.2, 11.6) more MVPA on weekdays versus weekend days, respectively. Children from mainland Europe had the largest differences (17.1 min/day more MVPA on weekdays versus weekend days, 95% CI: 15.3, 19.0) compared to the other countries/regions. Children who were classified as overweight/obese or normal weight/underweight accumulated 9.5 min/day (95% CI: 6.9, 12.2) and 10.9 min/day (95% CI: 8.3, 13.5) of additional MVPA on weekdays versus weekend days, respectively. Conclusions Children from multiple countries/regions accumulated significantly more MVPA on weekdays versus weekend days during school months. This finding aligns with the SDH and warrants future intervention studies to prioritize less-structured days, such as weekend days, and to consider providing opportunities for all children to access additional opportunities to be active.


2011 ◽  
Vol 14 (10) ◽  
pp. 1813-1822 ◽  
Author(s):  
Michela Barichella ◽  
Alexis E Malavazos ◽  
Giuseppe Fatati ◽  
Emanuele Cereda

AbstractObjectiveTo evaluate the awareness and knowledge about weight status and its management.DesignA 1 d cross-sectional survey. Basic anthropometric assessments (weight, height, BMI and waist circumference) and a self-administered questionnaire were considered.SettingNineteen Clinical Nutrition or Endocrinology and Metabolic Disorders Units or Dietetics Services in the Italian region of Lombardy.SubjectsAll adults attending the ‘Obesity Day’ initiative.ResultsA total of 914 participants (605 female and 309 male) were recruited. Although most of the participants (83·5 %) considered obesity to be a disease, 38·5 % were likely to misperceive their weight status. In particular, 38·8 % of normal-weight adults believed themselves to be overweight, whereas 71·1 % and 37·5 % of classes I and II/III obese adults classified themselves as being overweight and mildly obese, respectively. However, most of the overweight (90·2 %), mildly (96·8 %) and moderately/severely obese adults (99·1 %) recognized the need to lose weight. In all, 37·8 % of the sample underestimated the role of physical activity in weight management. Interestingly, only 17·2 % of dieters (previous or current) declared being advised by their doctor to lose weight. Multivariate models revealed that higher age, low education and higher BMI were important determinants of poor weight control and management. In addition, previous dieting appeared not to provide better knowledge, whereas the role of physical activity was recognized mainly by those practising it.ConclusionsThe present study suggests that in Italy knowledge about weight management should be improved not only in the general population but also among health-care professionals. To confirm this finding, there is now the rationale for a nationally representative survey. New educational programmes can be designed on the basis of the information collected.


2011 ◽  
Vol 8 (5) ◽  
pp. 682-692 ◽  
Author(s):  
Daniel Arvidsson ◽  
Mark Fitch ◽  
Mark L. Hudes ◽  
Catrine Tudor-Locke ◽  
Sharon E. Fleming

Background:Different movement efficiency in overweight children may affect accelerometer output data. The purpose was to investigate the ability of accelerometers to assess physical activity intensity and number of steps in normal-weight compared with overweight children.Methods:Eleven normal-weight and 14 overweight African American children walked at 2, 4, 5, and 6 km/h on a treadmill wearing Lifecorder, ActiGraph, RT3, and Biotrainer. Oxygen uptake was measured and steps manually counted. Fat free mass (FFM) was assessed from bioelectrical impedance analysis. Accelerometer counts and the individual linear regression lines of accelerometer counts versus VO2/FFM were evaluated, together with steps recorded by Lifecorder and Actigraph.Results:Correlations between accelerometer counts and VO2/FFM for all monitors were r ≥ .95 (P < .01). The accelerometer counts and their relationship to VO2/FFM did not generally differ significantly by body weight status. Lifecorder and Actigraph underestimated steps at 4, 5, and 6 km/h by less than 9%, but the error was up to −95% at 2 km/h.Conclusions:All 4 accelerometers show high ability to assess physical activity intensity, and can be used to compare physical activity between normal-weight and overweight children. The Lifecorder and the ActiGraph showed high accuracy in assessing steps, providing speed of movement exceeded 2 km/h.


Circulation ◽  
2014 ◽  
Vol 129 (suppl_1) ◽  
Author(s):  
Lisa Kakinami ◽  
Erin K O'Loughlin ◽  
Erika N Dugas ◽  
Catherine M Sabiston ◽  
Gilles Paradis ◽  
...  

Background: Compared to traditional non-active video games, exergaming may contribute significantly to overall daily physical activity (PA), but the association in population-based samples has not been fully explored. Objectives: To assess whether total minutes of PA per week among exergamers differs from non-exergamers, and whether the likelihood of meeting PA guidelines differs between exergamers and non-exergamers. Methods: Data were available from the 2011-2012 wave of the Nicotine Dependence in Teens (NDIT) study (n=829). Multivariable models assessed the association between exergaming with moderate, vigorous, and moderate-to-vigorous PA (MVPA) minutes in the past week, and the association between exergaming and meeting PA guidelines after adjusting for age, household income, employment status, education, weight status, and screen time. Exergaming was defined as use ≥ 1-3 times per month in the past year, and PA guidelines were defined as ≥ 150 minutes of moderately intense PA, or ≥ 75 minutes of vigorously intense PA, or a combination. Minutes of PA were measured using the well-validated International Physical Activity Questionnaire. Sex differences in exergaming use and exergaming beliefs and attitudes were also assessed. Results: In this population-based sample of young adults (mean age 24 [SD: 0.7] years, 55% female [n=455]), 18% (n=148) of the participants were exergamers (54% female, n=80). Exergaming males and females reported an average of 88.4 and 65.7 exergaming minutes in the past week, respectively. Compared to male exergamers, female exergamers were more likely to believe that exergaming is a good way to integrate PA into their lives (p<0.001), were more likely to prefer exergaming to indoor/outdoor sports (p=0.03), and were more likely to use fitness-training exergames at home (p=0.006). Perceived levels of exertion were also significantly different between male and female exergamers. The majority of male exergamers reported light (51%, n=23) or moderate (44%, n=20) exergaming exertion, whereas female exergamers reported moderate (53%, n=30), as well as light (28%, n=16) and intense (19%, n=11) exertion (p=0.02). After adjusting for covariates, male exergamers were not significantly different from male non-exergamers in total minutes of PA per week, but female exergamers reported 47 more minutes of moderate PA in the past week compared to female non-exergamers (p=0.03). There was no association between exergaming and meeting PA guidelines in either males or females. Conclusions: Exergaming contributes to moderate-intensity PA among females but not among males in this population based sample of young adults. Exergaming may therefore be an important source of PA for women, but the current landscape of exergames may be inadequate to substantially increase minutes of PA among males. Differences in attitudes toward exergaming should be further explored.


BMJ Open ◽  
2019 ◽  
Vol 9 (8) ◽  
pp. e025071 ◽  
Author(s):  
Liana Carmen Nagy ◽  
Muhammad Faisal ◽  
Maria Horne ◽  
Paul Collings ◽  
Sally Barber ◽  
...  

ObjectivesTo investigate factors associated with movement behaviours among White British (WB) and South Asian (SA) children aged 6–8 years during school terms and holidays.DesignCross-sectional.SettingThree primary schools from the Bradford area, UK.ParticipantsOne hundred and sixty WB and SA children aged 6–8 years.Primary and secondary outcomesSedentary behaviour (SB), light physical activity (LPA) and moderate-to-vigorous physical activity (MVPA) measured by accelerometry during summer, winter and spring and during school terms and school holidays. Data were analysed using multivariate mixed-effects multilevel modelling with robust SEs. Factors of interest were ethnicity, holiday/term, sex, socioeconomic status (SES), weight status, weekend/weekday and season.ResultsOne hundred and eight children (67.5%) provided 1157 valid days of data. Fifty-nine per cent of children were WB (n=64) and 41% (n=44) were SA. Boys spent more time in MVPA (11 min/day, p=0.013) compared with girls and SA children spent more time in SB (39 min, p=0.017) compared with WB children in adjusted models. Children living in higher SES areas were more sedentary (43 min, p=0.006) than children living in low SES areas. Children were more active during summer (15 min MVPA, p<0.001; 27 LPA, p<0.001) and spring (15 min MVPA, p=0.005; 38 min LPA, p<0.001) and less sedentary (−42 min and −53 min, p<0.001) compared with winter. Less time (8 min, p=0.012) was spent in LPA during school terms compared with school holidays. Children spent more time in MVPA (5 min, p=0.036) during weekend compared with weekdays. Overweight and obese children spent more time in LPA (21 min, p=0.021) than normal-weight children.ConclusionThe results of our study suggest that significant child level factors associated with movement behaviours are ethnicity, sex, weight-status and area SES. Significant temporal factors are weekends, school holidays and seasonality. Interventions to support health enhancing movement behaviours may need to be tailored around these factors.


2010 ◽  
Vol 35 (6) ◽  
pp. 826-833 ◽  
Author(s):  
Sarah J. Woodruff ◽  
Rhona M. Hanning

The purpose of this study was to determine diet quality and physical activity behaviours of grade 6 students by sex and body weight status, and to determine the associations between diet quality and physical activity behaviours. The Web-based Food Behaviour Questionnaire, which included a 24-h diet recall and the modified Physical Activity Questionnaire for Older Children (PAQ-C), was administered to a cross-section of schools (n = 405 students from 15 schools). Measured height and weight were used to calculate body mass index and weight status (Cole et al. 2000). A Canadian version of the Healthy Eating Index (HEI-C) was used to describe overall diet quality. The mean HEI-C was 69.6 (13.2) with the majority (72%) falling into the needs improvement category. The overall mean physical activity score was 3.7 out of a maximum of 5, with obese subjects being less active compared with normal weight and overweight (p < 0.001). Ordinal logistic regression analysis (of HEI-C vs. all measures of the PAQ-C, sex, and weight status) revealed that HEI-C ratings were likely to be higher in students that walked to and from school 5 days per week (vs. 0 days per week; odds ratio 3.18, p = 0.010); and were active 1 evening per week (vs. none; odds ratio 3.48, p = 0.039). The positive association between diet quality and some aspects of physical activity suggests possible clustering of health behaviours. Future research should test the potential benefits of promoting 1 health behaviour (e.g., healthy eating) with another (e.g., physical activity).


2015 ◽  
Vol 17 (2) ◽  
pp. 65-72 ◽  
Author(s):  
Robert W. Motl ◽  
Deirdre Dlugonski ◽  
Lara A. Pilutti ◽  
Rachel E. Klaren

Background: Behavioral interventions have significantly increased physical activity in people with multiple sclerosis (MS). Nevertheless, there has been interindividual variability in the pattern and magnitude of change. This study documented the efficacy and variability of a behavioral intervention for changing physical activity and examined the possibility that efficacy varied by the characteristics of individuals with MS. Methods: Eighty-two people with MS were randomly assigned to one of two conditions: behavioral intervention (n = 41) or waitlist control (n = 41). We collected information before the study on MS type, disability status, weight status based on body-mass index, and current medications. Furthermore, all participants completed the Godin Leisure Time Exercise Questionnaire and the abbreviated International Physical Activity Questionnaire and wore an accelerometer for 1 week to measure minutes of moderate-to-vigorous physical activity before and after the 6-month intervention period. Results: Analysis of covariance (ANCOVA) indicated that participants in the behavioral intervention had significantly higher levels of physical activity than control participants after the 6-month period (P &lt; .001). There was substantial interindividual variability in the magnitude of change, and ANCOVA indicated that MS type (relapsing vs. progressive) (P &lt; .01), disability status (mild vs. moderate) (P &lt; .01), and weight status (normal weight vs. overweight/obese) (P &lt; .05) moderated the efficacy of the behavioral intervention. Conclusions: The behavioral intervention was associated with improvements in physical activity, particularly for those with mild disability, relapsing-remitting MS, or normal weight status.


Author(s):  
Ana Gutiérrez-Hervás ◽  
Ernesto Cortés-Castell ◽  
Mercedes Juste-Ruíz ◽  
Antonio Palazón-Bru ◽  
Vicente Gil-Guillén ◽  
...  

Introduction.  Interpretation of accelerometer-derived physical activity in preschool children is confounded by differences in cut-off points.Aim. The purpose of this study was to analyze physical activity in 2-to-7-year-old children to establish reference values for daily activity.   Methods.  Observational study in children aged 2-7 years, without chronic diseases and whose parents provided informed consent. The main variable was physical activity, measured continuously over 120 hours (three workdays and two weekend days) by accelerometer. Secondary variables were weight status (BMI Z-score) and gender. The relationship between the main variable and secondary variables was determined through the t-test, ANOVA and the Pearson correlation coefficient.  A multivariate model was used to obtain the standard deviation of all possible combinations of values, constructing percentiles of normality(x±s and x±2·s).Results. 136 children (35% of municipality children), 54.4% girls. Their weight status distribution was: 25 underweight(18.4%), 35 normal weight(25.7%), 40 overweight(29.4%) and 36 obese(26.5%). The median age was 5.7 years and the mean physical activity was 591.9 counts/minute. The boys undertook more physical activity(p=0.031) and the underweight and normal-weight children undertook more physical activity than the overweight and obese children(p=0.032). There were no significant differences according to age.  The multivariate analysis showed significant differences(p<0.001) according to gender and weight status. In boys, physical activity decreased as weight status increased. In contrast, the girls in the extreme BMI groups obtained higher levels of physical activity.Conclusion.  The reference values obtained, categorized by gender and weight status, provide clinicians with standardized daily physical activity levels of preschool-age children.


2019 ◽  
Author(s):  
Stella T. Lartey ◽  
Costan G. Magnussen ◽  
Lei Si ◽  
Godfred O. Boateng ◽  
Barbara de Graaff ◽  
...  

AbstractBackgroundStudies on changes in the prevalence and determinants of obesity in older adults living in sub-Saharan Africa are limited. We examined recent changes in obesity prevalence and associated factors for older adults in Ghana between 2007/08 and 2014/15.MethodsData on adults aged 50 years and older in Ghana were collected in the WHO SAGE Wave 1 (2007/08; n=4158) and Wave 2 (2014/15; n=1663). The weighted prevalence of obesity, overweight, normal weight and underweight, and of high central adiposity was compared in Waves 1 and 2. Multinomial and binomial logistic regressions were used to examine whether the determinants of weight status based on objectively measured body mass index and waist circumference changed between the two survey periods.ResultsThe prevalence of obesity (Wave 1=10.2%, 95% CI: 8.9-11.7%; Wave 2=15.0%, 95% CI: 12.6-17.7%) and overweight (Wave 1=19.6%, 95% CI: 18.0-21.4%; Wave 2=24.5%, 95% CI: 21.7-27.5%) was higher in Wave 2 than Wave 1 and more than half of the population had high central adiposity (Wave 1=57.7%, 95% CI: 55.4-60.1%; Wave 2=66.9%, 95% CI: 63.7-70.0%) in both study periods. Obesity prevalence was 16% lower in males and 55% higher in females comparing Wave 1 to Wave 2. Female sex, urban residence, and high household wealth were associated with higher odds of overweight/obesity and high central adiposity. Those aged 70+ years had lower odds of obesity in both study waves. In Wave 2, females with physical activity level were more likely to be obese.ConclusionOver the 7-year period between survey waves, the population prevalence of overweight and obesity increased by 25% and 47%, respectively, while underweight reduced by 43%. These findings differed considerably by sex, which points to differential impacts of past initiatives to reduce overweight/obesity, potential high-risk groups in Ghana, and the need to increase surveillance.


Sign in / Sign up

Export Citation Format

Share Document