Latent Class Analysis of Lifestyle Risk Factors and Association with Overweight and/or Obesity in Children and Adolescents: Systematic Review

Abstract

Background:

Multiple modifiable lifestyle factors are well-known contributors to many health problems.

Objectives:

This study aims to determine the association between latent class analysis (LCA) of modifiable lifestyle risk factors with being overweight and/or obese for children and/or adolescents.

Methods:

Articles were selected from six databases, without limitation regarding language or date. The review included studies that identify latent classes of modifiable lifestyle risk factors [e.g., physical activity (PA), diet, sedentary behavior (SB), and/or unhealthy behavior] by LCA to determine the association between latent classes with being overweight and/or obese. The methodology of the selected studies was evaluated using the JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies. Owing to the heterogeneity between latent classes of modifiable lifestyle risk factors with obesity and/or being overweight, the results are described narratively.

Results:

Using a selection process in two phases, nine articles were included. All of the included studies were of high methodological quality. The studies were conducted in six different countries: the USA, Brazil, Canada, Portugal, Italy, and Australia. Sample sizes ranged from 166 to 18.587 children and adolescents, and in terms of age (range 5–19 years). Across study clusters characterized by low consumption of fruit and vegetables, and high consumption of fatty foods, sugar snack foods, sweets, chips and fries, low PA (<1 hour each day), and high SB (screen time and TV >2 hours/day), sleep time (<10 hours/day) were positively associated with being overweight and/or obese.

Conclusion:

Overall there is good evidence to support that the modifiable lifestyle risk factors clustered together by LCA should be novel targets for the treatment of obesity and its associated comorbidities.

Introduction

Being overweight and/or obesity (Ow/Ob) among children and adolescents are major public health problems that have increased in prevalence in almost all countries since 1990,^1,2 with the WHO reporting 41 million overweight children aged under five in 2016 worldwide.³ Ow/Ob in childhood are predictors of the risk of obesity, morbidity, and mortality in adulthood^4–7 and are associated with the development of various comorbidities⁸ (e.g., dyslipidemia, nonalcoholic steatohepatitis, type 2 diabetes mellitus, obstructive sleep apnea, hypertension, and polycystic ovary syndrome), as well as psychosocial problems, including discrimination, social isolation, and low self-esteem, which can themselves affect health, education, and quality of life.⁸

To prevent comorbidities and morbidity in adult life, it is necessary to intervene to reduce obesity⁹ through the joint analysis of lifestyle behaviors and their interrelationships,^10–13 especially physical activity (PA), sedentary lifestyles, and dietary behavior.^1,14 Socioeconomic status, age, gender, ethnicity, sleep duration, depression, symptoms of anxiety, and behavioral problems should be taken into account,^9,15,16 as they may be associated with the risk of developing obesity. In addition, parenting style and parental support,¹⁷ low levels of parental education,¹¹ social inequality and gender differences,¹⁸ types of friendships, adolescent sex, and race/ethnicity should also be considered.¹⁹

Some statistical tools, such as cluster analysis (CA), factorial analysis, and latent class analysis (LCA), are used to classify lifestyle factors as more or less healthy.^13,20 LCA has advantages over the other methods due to using an unobserved (latent) variable based on the responses of a set of observed variables (continuous, nominal, and ordinal), grouping homogeneous and mutually exclusive individuals (cases and units) into classes.^13,21–25

LCA identifies the lowest number of latent classes and then associates the observed symptoms, for example, lifestyle factors, with behavioral patterns that exist within a heterogeneous population.^26,27 The different health behaviors often cannot be directly measured; that is, they are latent variables and the LCA serves as an ideal analytical method because it takes into account the observable variables and the unobservable variables (latent), with the objective of identifying classes with homogeneous behaviors.

LCA has been used in several studies to explore modifiable lifestyle risk factors associated with being Ow/Ob, such as diet, PA, substance use, sexual behavior, stress, sleep,^23,28 neurocognitive processes associated with self-regulation, decision-making, goal-directed behavior,²⁹ and asthma phenotypes.³⁰ LCA has been used with other modifiable lifestyle risk factors, such as alcohol use,³¹ risk factors for suicide,^23,27,28 and impulsivity, anxiety, depression, drug use, and alcohol dependence.³²

Thus, knowledge of lifestyle risk factors through LCA is relevant, as it can indicate whether the risk factors together have synergistic effects and can negatively affect health, thereby supporting preventive health interventions. LCA can also be part of implementing strategies that will maximize the impact of public health.²² We hypothesized (1) that LCA would identify at least one pattern characterized by obesogenic behaviors (unhealthy food, physical inactivity, or a sedentary lifestyle) and (2) that these isolated or clustered patterns may be associated with the development of Ow/Ob in children and adolescents.

Therefore, the aim of this study was to identify latent classes of obesogenic behaviors in children/adolescents based on the clustering of the modifiable lifestyle risk factors of PA, diet, sedentary behavior and/or unhealthy behavior (SB/UB), and then determine the association between these latent classes and Ow/Ob.

Methods

Protocol and Registration

This systematic review was reported following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) checklist.³³ The protocol underpinning this review is available through the PROSPERO database (see registration no. CRD42019132083).

Eligibility Criteria

Inclusion criteria

For inclusion, studies had to meet the following requirements: 1.

Study design: Nonexperimental designs (observational, cohort studies, surveys, cross-sectional, and longitudinal) were considered.

Population: Studies with samples of children aged <11 years, and/or adolescents aged 12–18 years (or a mean within these ranges) at baseline were included. Information was reported by the children/adolescents aged >12 years or their parents, mother–child pairs, participating families, or parents or caregivers of children <12 years.

Comparison: Studies could target multiple behaviors in any combination.

Language and date: Studies that were published in any language or date were considered.

Types of measurement tools: Studies that identified latent classes of modifiable lifestyle risk factors (e.g., diet, PA, and SB/UB) by LCA and estimation [item-response probabilities, greater likelihood >0.50, or probability (%)].³⁴

Outcome:

Primary outcomes: We included studies reporting LCA based on the clustering of modifiable lifestyle risk factors: diet, PA, and SB/UB.

Secondary outcomes: We included studies that determined the association between latent classes with Ow/Ob. Variety of obesity measures: body weight status was determined using age- and gender-specific BMI percentiles (Ow BMI ≥85th percentile to 95th percentiles, Ob 95th percentile), BMI z-scores, BMI standard deviation scores (SDSs) (Ow one standard deviation above the average and Ob two standard deviations above the average), BMI cut-off points. BMI was calculated as weight in kilogram divided by the square of height in meters (BMI = weight (kg)/(height (meter))²).³⁵

The primary criteria used to define being Ow/Ob include overweight: BMI ≥85th percentile to 95th percentile, BMI = one standard deviation above the average; obesity: BMI = 95th percentile, BMI = two standard deviations above the average, and considered overweight if their BMI z-score was between the 85th and 95th percentile and obese if their BMI z-score was above the 95th percentile.³⁵

Supplementary Appendix Table SA1 contains modifiable lifestyle risk factors, variables, used as the latent class indicators.

Exclusion criteria

Experimental designs (e.g., randomized controlled trial [RCT], interrupted time series, and pretest–post-test)

Studies with samples consisting of adults, elderly, pregnant, and nonhealthy subjects (people with type 2 diabetes, hypertension patients, and eating disorders)

Studies that do not use the LCA based on the clustering of modifiable lifestyle risk factors: diet, PA, and SB/UB

Studies that do not use the LCA and do not use estimation [item-response probabilities, greater likelihood >0.50, or probability (%)]

Studies that do not correlate modifiable lifestyle risk factors with obesity parameters

Studies that do not identify at least three latent classes of modifiable obesogenic behaviors in lifestyle risk factors diet, PA, and SB/UB.

Supplementary Appendix Table SA2 contains the articles excluded and the reasons for exclusion.

Information Sources and Search Strategy

Detailed individual search strategies for each of the following electronic databases were performed: Cochrane Library, Lilacs, Eric, Livivo, Pubmed/Medline, and Social Care Online databases. We performed a gray literature search on Google Scholar. In addition to the electronic search, a hand search was made, and reference lists of the selected articles were screened.

Search terms and Boolean operators with the list of medical subject headings (MeSH), keywords, and/or other controlled vocabulary terms used in searches of the PubMed and the other electronic databases were used. Database-specific search terms included the keywords “latent class analysis,” “obesity,” “childhood,” “children,” “adolescents,” “diet,” and relevant synonyms. All electronic searches were conducted from their earliest records up to March 1, 2019 and updated until May 1, 2019.

Supplementary Appendix Table SA3 contains the search terms, search strategy, and databases.

Study Selection

We selected the articles in two phases. In phase l, two authors (R.L. and F.D.C.) independently reviewed the titles and abstracts for all of the references. During this phase, any articles that did not meet the eligibility criteria described earlier in the inclusion and exclusion criteria section were excluded. In phase 2, they applied the same selection criteria to the full text of the articles to determine which ones to include. The same two authors independently reviewed all full-text articles. When they needed help making a final decision, a third author (E.K. or M.A.A.A.) participated in the review process. Supplementary Appendix Table SA2 contains the articles excluded and the reasons for exclusion.

Data Collection Process and Data Items

Two authors (R.L., F.D.C.) extracted data from the selected studies, including author, year of publication, location, and information reported; population, sample size characteristics (boys, girls, and total), and weight status prevalence; LCA based on the clustering of modifiable lifestyle risk factors; findings variables (latent class indicators) and main conclusion. A third author (E.K.) cross-checked all the retrieved information.

Risk of Bias in Individual Studies

Risk of bias in individual studies included articles was conducted by the JBI-MAStARI (JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies). The questionnaire consists of eight questions that were answered with yes, no, unclear, or not applicable.³⁶ In the JBI-MAStARI, two criteria listed in the instrument were deemed irrelevant to the nature of the studies. We classified the studies as follows: high methodological quality (>5 times a score of “yes”), moderate methodological quality (3–4 times a score of “yes”), or low methodological quality (0–2 times a score of “yes”).³⁷ Supplementary Appendix Table SA4 contains the risk of bias of the selected studies by JBI-MAStARI.

Summary Measures

We considered as outcome of the association modifiable lifestyle risk factors (diet, PA, and SB/UB) identified by LCA and association with Ow/Ob. Variables used as the latent class indicators, identified by item-response probabilities.

Synthesis of Results

Owing to heterogeneity relating between assessment of associations of modifiable lifestyle risk factors (diet, PA, and SB/UB) and BMI, it was not possible to perform a meta-analysis, and thus results are described narratively. It was not possible to assess publication bias through funnel plots, as no statistical data synthesis could be performed.

Results

Study Selection

Of the 1114 identified articles, 9 were included and the results from these studies were weighted. A flowchart of the process of study identification and selection is shown in Figure 1.

Figure 1.

Flowchart of the literature search and selection criteria.

Study Characteristics

All the studies were published between 2011 and 2019. No language restriction was set, although all the articles included were published in English. The studies collected information on modifiable lifestyle risk factors from children, adolescents, and their parents.

The studies were conducted in six different countries, with the majority being from North and South America (USA,^38,40–42 Brazil,⁴⁶ and Canada⁴⁴), two from Europe (Portugal⁴³ and Italy⁴⁵) and one from Australia.³⁹

Sample sizes ranged from 166⁴² to 18,587⁴⁴ children and adolescents. With the exception of one study³⁹ that included children aged 6–7 years, the ages of the children and adolescents in these studies ranged from 5 to 19 years. The sample sizes were moderate to large, with one study using a sample of <500,⁴² four studies using samples between 500 and 1000,^38,41,43,45 and four studies including >1000 participants.^39,40,44,46

The total sample from the nine selected studies included 38,512 children and adolescents (19,419 boys and 19,093 girls). The samples of eight studies comprised boys and girls, and one study exclusively consisted of girls.⁴² The total sample of the American countries was n = 35,111,^{38,40–42,44,46} of the European countries was n = 1569,^43,45 and n = 1832 for Australia.³⁹ A summary of the descriptive characteristics of the included studies is provided in Table 1.

Table 1.

Summary of the Characteristics of the Included Studies

Author, year, location study design	Population and sample/weight status prevalence	LCA based on the clustering of modifiable lifestyle risk factors	Variables, used as the latent class indicators and their key covariates = classIdentified and item-response probabilities (showed greater likelihood >0.00 or probability [%])	Conclusion class = (%)
Huh et al. (2011)³⁸ Southern California, USA Information reported by children and their parents	997 out of 1005 children 9.26 ± 0.48 years (518 boys and 479 girls) (Ow/Ob 406 and 591 normal)	Class 1 (15.9%) high-sedentary, HF/HS snacks, weight conscious Class 2 (24.9%) dieting without exercise, weight conscious Class 3 (13.3%) high-sedentary, HF/HS snacks, not weight conscious Class 4 (13.7%) active, healthy eating Class 5 (32.2%) low healthy and snack food, inactive, not weight conscious	(a) Diet (≥daily consumption) (a1) Salty high-fat snack food = 1 (0.71), 2 (0.17), 3 (0.83), 4 (0.23), 5 (0.23) (a2) High-sugar snack food = 1 (0.71), 2 (0.06), 3 (0.83), 4 (0.13), 5 (0.13) (a3) Any fruits = 1 (0.29), 2 (0.17), 3 (0.22), 4 (0.91), 5 (0.15) (a4) Any vegetable = 1 (0.29), 2 (0.17), 3 (0.22), 4 (0.85), 5 (0.15) (b) PA (b1) Daily, 30+ minutes = 1 (0.22), 2 (0.23), 3 (0.24), 4 (0.54), 5 (0.27) (b2) Team sports = 1 (0.69), 2 (0.67), 3 (0.64), 4 (0.81), 5 (0.70) (c) SB/UB (c1) Screen time (2+ hours/day) = 1 (0.86), 2 (0.38), 3 (0.88), 4 (0.33), 5 (0.48)	Normal/Ow/Ob (BMI percentiles) 1 (38)/(62) 2 (28)/(72) 3 (82.6)/(17.4) 4 (61.9)/(38.2) 5 (85.8)/(14.2)
Magee et al. (2013)³⁹ Australia Information reported by children and their parents	1833 children 6–7 years (945 boys and 887 girls) (Ow/Ob 334 and 1.498 normal)	Class 1 (27.7%) healthy Class 2 (24.8%) sedentary Class 3 (13.3%) unhealthy eaters	(a) Diet (times/day) (a1) Fruit/vegetable (<2) = 1 (4.5%), 2 (45.7%), 3 (19.8%) (a2) Fruit/vegetable (3–4) = 1 (46.1%), 2 (45.1%), 3 (59.2%) (a3) Fruit/vegetable (>4) = 1 (49.4%), 2 (9.2%), 3 (21%) (a4) High fat food (<1) = 1 (54.3%), 2 (13.4%), 3 (6.9%) (a5) High fat food (1) = 1 (44.1%), 2 (48.1%), 3 (43.9%) (a6) High fat food (<2) = 1 (1.6%), 2 (38.5%), 3 (49.2%) (a7) High sugar drinks (0) = 1 (8.4%), 2 (29.7%), 3 (37.6%) (a8) High sugar drinks (>1) = 1 (15.6%), 2 (70.3%), 3 (62.4%) (b) PA (each day) (b1) <1 hour = 1 (26%), 2 (80.7%), 3 (0.0) (b2) >2 hours = 1 (37.4%), 2(0.0%), 3 (58.5) (c) SB/UB (each day) (c1) Screen time (<1 hour) = 1 (58.1%), 2 (11.4), 3 (0.0) (c2) Screen time (>2 hours) = 1 (7.7%), 2 (52.3%), 3 (40.7%) (c3) sleep duration (>11 hours) = 1 (52.2%), 2 (36.6%), 3 (22.4%)	Normal/Ow/Ob (BMI cutoff points) 1 (84.4)/(15.6) 2 (76.9)/(23.1) 3 (82.8)/(17.2)
Iannotti and Wang (2013)⁴⁰ USA Information reported by adolescents	9.206 out of 9.227 Mean age: 13.9 years (4.456 boys and 4.750 girls) (Ow/Ob 2.826 and 6.380 normal)	Class 1 (26.5%) healthful Class 2 (26.4%) unhealthful Class 3 (47.2%) typical	(a) Diet (a1) Any fruits = 1 (83%), 2 (45.5%), 3 (10.1%) (a2) Any vegetable = 1 (79.3%), 2 (35.5%), 3 (9.4%) (a3) Intake of sweets = 1 (22%), 2 (73.3%), 3 (16.3%) (a4) Sweetened soft drinks = 1 (10.7%), 2 (77.8%), 3 (24.5%) (a5) Chips = 1 (8.5%), 2 (80.9%), 3 (7.5%) (a6) Fries = 1 (2.8%, 2 (44.4%), 2 (2.7%) (b) PA (b1) 60 minutes/day at least 5 days/week = 1 (63.9%), 2 (50%), 3 (45.6%) (c) SB/UB (c1) Exceeding 2 hours/day TV = 1 (37.7%), 2 (74.5%), 3 (53.5%)	Normal/Ow/Ob (BMI percentiles) 1 (71.8)/(28.2) 2 (70.1)/(29.9) 3 (67.2)/(32.8)
Wang et al. (2014)⁴¹ Midwestern city Nebraska/USA Information reported by adolescents	791 out of 894 12–18 years (419 boys and 372 girls) (Ow 190 and 601 normal)	Class 1 (52%) healthy dietary behavior and PA, less permissive parenting, and physician advice Class 2 (30%) moderately healthy dietary behavior and PA, less permissive parenting, no physician advice Class 3 (18%) unhealthy dietary behavior and PA, more permissive parenting, and physician advice	(a) Diet (a1) Eats breakfast every day = 1 (0.76), 2 (0.55), 3 (0.32) (a2) Eats 5 servings of fruits and vegetables on 4 or more days a week = 1 (0.76), 2 (0.54), 3 (0.45) (b) PA (b1) 60 minutes on 4 or more days a week = 1 (0.90), 2 (0.80), 3 (0.56) (b2) Participates on at least 1 sports team = 1 (0.92), 2 (0.73), 3 (0.47) (c) SB/UB (c1) Watches TV or plays vídeo games <2 hours/day = 1 (0.78), 2 (0.63), 3 (0.42)	Normal/Ow/Ob (BMI percentiles) 1 (79)/(21) 2 (80)/(20) 3 (59)/(41)
Balantekin et al. (2015)⁴² Pennsylvania, USA Information reported by adolescents and their parents	166 girls out of 197 families 5–15 years (Ow/Ob 31 and 135 normal)	Class 1 (26%) nondieters Class 2 (16%) lifestyle Class 3 (43%) dieters Class 4 (17%) extreme dieters	(a) Diet (a1) Fruits and vegetables = 1 (0.03), 2 (0.94), 3 (0.85), 4 (1.00) (a2) Eat less fat = 1 (0.02), 2 (0.34), 3 (0.90), 4 (1.00) (a3) Eliminate snacking, sweets, junk = 1 (0.01), 2 (0.47), 3 (0.93), 4 (0.99) (a4) Reduce calories and amount of food = 1 (0.09), 2 (0.15), 3 (0.87), 4 (0.97) (a5) Eat low calorie food = 1 (0.02), 2 (0.11), 3 (0.62), 4 (0.83) (a6) Eat less meat = 1 (0.05), 2 (0.00), 3 (0.08), 4 (0.90) (a7) Skip meals = 1 (0.12), 2 (0.03), 3 (0.54), 4 (0.94) (b) PA (b1) Increase exercise = 1 (0.26), 2 (0.69), 3 (0.98), 4 (0.84) (c) SB/UB (c1) Unhealthy behavior (use of laxatives/enemas, diuretics, diet pills, or appetite suppressants, smoking cigarettes, or vomiting) = 1 (0.02), 2 (0.00), 3 (0.11), 4 (0.54)	Normal/Ow/Ob (BMI percentiles) 1 (88)/(12) 2 (92)/(8) 3 (80)/(20) 4 (62)/(38)
Pereira et al. (2015)⁴³ Portugal, Europe Information reported by children and their parents	686 children 9–11 years (305 boys and 381 girls) (Ow/Ob 314 and 372 normal)	Class 1 (35.3%) sedentary, poorer diet quality Class 2 (64.7%) insufficiently active, better diet quality	(a) Diet (a1) Fruits/vegetables (<7 days) = 1 (78.1%), 2 (67.6%) (b) PA (b1) <60 minutes of daily = 1 (52.1%), 2 (69.8%) (c) SB/UB (c1) Sleep time (<10 hours·day −1) = 1 (90.1%), 2 (94.1%) (c2) Screen time (≥120 minutes·day −1) = 1 (71.1%), 2 (59.2%)	Normal/Ow/Ob (BMI cutoff points) 1 (59.1)/(40.9) 2 (51.6)/(48.4)
Laxer et al. (2017)⁴⁴ Ontario and Alberta/Canadá	18.587 out of 30.147 students 14–18 years (9.498 boys and 9.089 girls) (Ow/Ob 5.093 and 13.494 normal)	Class 1 (24%) traditional school athletes Class 2 (43.4%) inactive screenagers Class 3 (16%) health conscious Class 4 (16.6%) moderately active substance users	(a) Diet (a1) Fast food consumption = 1 (81.5%), 4 (85%) (a2) Breakfast daily = 3 (79.5%) (a3) Refraining from fast food = 3 (71.6%) (a4) Refraining other snack = 3 (93.6%) (a5) Refraining sugar-sweetened beverage consumption = 3 (93.7%) (a6) Snacks from off-school property = 4 (52%) (a7) Sugar-sweetened beverages = 4 (47%) (b) PA (b1) 60 minutes of daily = 1 (64%), 2 (37%), (b2) Participating in intramural = 1 (87%), 2 (11%), 3(62%), (b3) Varsity sports = 1 (99%), 2 (8%), 3 (73%) (b4) Strength training at least three times per week = 2 (26%), 3 (59%) (c) SB/UB (c1) 2 or more hours watching TV = 2 (59%), 4 (57%) (c2) Surfing the internet = 2 (58%), 4 (57%) (c3) Tobacco users = 4 (70%) (c4) Binge drinkers = 1 (26%), 4 (79%) (c5) Marijuana users = 1 (9.5%),4 (83%) (c6) Playing video games = 2 (35%).	Normal/Ow/Ob (BMI cutoff points) 1 (74.4)/(25.6) 2 (73.9)/(26.1) 3 (78.9)/(21.1) 4 (71.7)/(28.3)
Tabacchi et al. (2018)⁴⁵ Palermo, Italy Information reported by adolescents and their parents	883 adolescents 13–19 years (550 boys and 333 girls) (Ow/Ob 248 and 635 normal)	Class 1 (30.7%) virtuous Class 2 (18.8%) low PA/sport Class 3 (25.8%) incorrect alcohol/foods habits Class 4 (15.9%) health risk/overweight Class 5 (8.8%) malaise/diseases	(a) Diet (a1) Nonadequate meals = 1 (0.35), 2 (0.60), 3 (0.93), 4 (0.43), 5 (0.45) (a2) Nonadequate food habits = 1 (0.31), 2 (0.45), 3 (0.75), 4 (0.40), 5 (0.45) (b) PA (b1) <60 minutes/day = 1 (0.0), 2 (0.33), 3 (0.01), 4 (0.0), 5 (0.04) (b2) Sport <3 hours/week = 1 (0.27), 2 (0.46), 3 (0.50), 4 (0.49), 5 (0.51) (c) SB/UB (c1) Screen watching >2 hours/day = 1 (0.27), 2 (0.46), 3 (0.50), 4 (0.50), 5 (0.51) (c2) Alcohol consumption = 1 (0.57), 2 (0.55), 3 (0.76), 4 (0.65), 5 (0.80) (c3) Smoking = 1 (0.18), 2 (0.09), 3 (0.13), 4 (0.09), 5 (0.16)	Normal/Ow/Ob (BMI percentiles—asso. project) 1 (87)/(13) 2 (69)/(31) 3 (96)/(4) 4 (20)/(80) 5 (60)/(40)
Lobo et al. (2019)⁴⁶ Florianopolis, Brazil Information reported by adolescents and their parents	5.364 out of 9200 children 7–12 years (2.728 boys and 2.636 girls) (Ow/Ob 2.001 and 3.363 normal)	Class 1 (41.3%) traditional Class 2 (36.3%) monotonous Class 3 (22.4%) mixed	(a) Diet (a1) Beans (cooked) = 1 (43%) (a2) Pizza/hamburger/hot-dog = 2 (49%) (a3) Maize/potatoes/breakfast cereals = 3 (122%) (a4) Cheese bread = 3 (181%) (a5) French fries = 3 (195%) (a6) Chips = 3 (186%) (a7) Nuggets = 3 (171%) (a8) Candies/chocolate/lollypops/ice cream = 3 (104%) (a9) Soda = 3 (96%) (a10) Pasta = 2 (23%) (b) PA+c) SB/UB^a Boys (b1) Lowest PA and lowest SB = 1 (36.9%), 2 (52.2%), 3 (10.6%) Girls (b2) Lowest PA and lowest SB = 1 (46.6%), 2 (41.8%), 3 (11.6%) (b3) Lowest PA and intermediary SB = 1 (51.1%), 2 (40.2%), 3 (8.7%)	Normal/Ow/Ob (BMI z-scores) 1 (56.8)/(43.2) 2 (57)/(43) 3 (79)/(21)

BMI was computed using the standard formula [body mass (kg)/height (m)²]. Height and weight were self-reported and used to calculate BMI (kg/m²). Overweight and obesity were defined at cutpoints of 25 and 30 kg/m², respectively. Or children were classified as overweight if their BMI z-score was between the 85th–95th percentile and obese if their BMI z-score was above the 95th percentile.

SB = board games, playing with dolls, playing with toy cars, watching TV, listening to music, using smartphone/tablet, using computer, play videogames.

HF/HS snacks, high-fat and/or high-sugar snacking; LCA, latent classes analysis; Ow/Ob, overweight and/or obesity; PA, physical activity; SA, screen-based sedentary activities self-reports; SB/UB, sedentary behavior and/or unhealthy behavior.

Brief food frequency questionnaires^38,40,43 were among the most commonly used methods to assess individual long-term dietary intake of foods and nutrients in the studies. These are appropriate for investigating dietary patterns on the basis of frequencies. Dietary intake data were most commonly collected using a one 24-hour,³⁹ questions adapted from the CDC's,⁴¹ 24 hours-Drecall—3,⁴² Canada's Food Guide serving sizes,⁴⁴ ASSO-FHQ (Food Habits Questionnaire),⁴⁵ and Web-CAAFE (Food Intake and Physical Activity of Schoolchildren).⁴⁶

The type, duration, and frequency of the PA intervention varied among the studies. Studies focused on daily hard activities lasting >30 or ±60 minutes per day,^{38–41,43–46} traditional sports,³⁸ fitness activities and other sports,⁴⁵ physical activities, varsity sports, and intramural.⁴⁴ Of these, studies used an objective measure: Questions adapted from CDC's,^38,40,41 Australian guidelines,³⁹ moderate to vigorous physical activity (MVPA) daily recommendations,⁴³ Canada's Physical Activity Guidelines for Children and Youth,⁴⁴ ASSO-PASAQ (Physical Activity, Smoke, and Alcohol Questionnaire)⁴⁵ and Web-CAAFE.⁴⁶

Studies used an objective measure for SB/UB: The American Academy of Pediatrics (AAP),^38,40 Australian guidelines,³⁹ Questions adapted from the CDC,^41,42 Recommendations by the National Sleep Foundation and screen time by questionnaire,⁴³ Canadian Sedentary Behaviour Guidelines,⁴⁴ ASSO-PASAQ,⁴⁵ and Web-CAAFE.⁴⁶

Supplementary Appendix Table SA5 contains the methods to assess PA, diet, and SB/UB.

Primary Outcome: LCA Based on the Clustering of Modifiable Lifestyle Risk Factors (Diet, PA, and SB/UB)

Table 1 presents the nine studies that examined the relationships among modifiable lifestyle risk factors derived from LCA.

LCA involves testing sequential models to identify the optimal number of latent classes³⁹ (Supplementary Appendix Table SA6). The number of clusters from LCA observed across the studies varied among two,⁴³ three,^39–41,46 four,^42,44 and five.^66,73

All the studies examined associations between modifiable lifestyle risk factor outcomes for boys and girls combined, whereas one examined the association between PA and SB separately for boys and girls.⁴⁶

Diet correlates of clusters

All the studies indicated the frequency with which each child or adolescent consumed several different types of foods per day (≥daily consumption).

Dietary behavior (what the children or adolescents consumed) was measured: in Australia³⁹ (consumption of fruits and vegetables, high-fat food, and high-sugar drinks); in Portugal⁴³ (consumption of fruits and vegetables); in Italy⁴⁵ (inadequate meals and inadequate food habits); in Canada⁴⁴ (consumption of fast food, breakfast, snacks, and sugar); in Brazil⁴⁶ (consumption of beans, pizza/hamburgers/hot dogs, maize/potatoes/breakfast cereals, cheese bread, fries/nuggets, and sugar/candies/chocolate/lollipops/ice cream; in the USA^38,40–42 (consumption of fruits and vegetables, fatty foods, sugar/sweets, fries/chips, and breakfast).

Six studies analyzed the consumption of fruit and vegetables designated: Healthy eating,³⁸ Healthy,^39,40 Healthy dietary behaviour,⁴¹ Lifestyle, dieting and extreme dieting⁴² and better diet quality.⁴³ The practice of eating breakfast every day was analyzed in two studies and the classes named were healthy dietary behaviour⁴¹ and health conscious.⁴⁴

The classes designated: high-sedentary, high-fat and/or high-sugar snacks, not weight conscious,³⁸ sedentary,³⁹ unhealthy eaters,³⁹ unhealthful,⁴⁰ nondieters,⁴² traditional school athletes,⁴⁴ inactive screenagers,⁴⁴ and mixed⁴⁶ involved the increased consumption of obesogenic foods, including intake of snacks, sweets, junk food, fast food, chips, candies, chocolate, ice cream, pizza, hamburgers, hot dogs, salty high-fat snack foods, high-sugar snack foods, and sugary drinks, among others.

PA correlates of clusters

All the studies indicated the frequency with each children or adolescents engaged in at least one kind of exercise, PA, participated on a sports team, varsity sport, or intramural activities per day.

PA was measured (whether the children or adolescents engaged in at least the following): in Australia³⁹ (one PA per day); in Portugal and in Italy: (moderate PA each day,^43,45 and participated on at least one sports team³⁹); in USA, Brazil, and Canada: (practiced PA each day,^{38–42,44,46} participated in intramural activities,⁴⁴ varsity sports,⁴⁴ participated on at least one sports team,^38,41 and performed strength training at least three times per week³⁹).

The classes designated: active,³⁸ healthy,³⁹ healthful,⁴⁰ healthy,⁴¹ and lifestyle⁴² involved 1- and 2-hour physical activities per day, and 4 or more days per week. And the classes designated: high-sedentary,³⁸ sedentary,³⁹ typical,⁴¹ unhealthy PA,⁴² sedentary and insufficiently active,⁷¹ involved <1 hour PA each day.

SB/UB correlates of clusters

All the studies indicated the number of hours each children or adolescent spent in SB/UB per day. SB/UB included analysis by (1) weight perception, related behaviors, weight-loss behavior⁴⁴; (2) screen time/watching/streaming TV^{38–41,43–46}; (3) playing video or computer games^40,46; (4) using the internet/computer (surfing the internet)⁴⁴; (5) sleep time^43–45; (6) use of laxatives/enemas, diuretics, diet pills, or appetite suppressants, smoking cigarettes, tobacco, marijuana use, or self-induced vomiting⁴²; (7) listening to music, using smartphone/tablet.⁴⁶

Secondary Outcome: Association between Modifiable Lifestyle Risk Factors with Childhood Ow/Ob Risk

Table 1 summarizes the nine studies identified that have examined associations between the cluster of modifiable lifestyle risk factors and weight status or BMI.

In all the studies, children and adolescents were classified as Ow/Ob according to the BMI cutoff points,^43,44 BMI z-score,⁴⁶ or BMI percentiles.^{38,40–42,45} The total sample of the nine selected studies included 38,512 children and adolescents: 27,069 (70.28%) classified as normal and 11,443 (29.72%) as Ow/Ob.

Across five study clusters characterized by low consumption of fruit and vegetables^38–41,43 and high consumption of fatty foods, sugary snack foods, sweets, chips and fries,^38–40 low PA (<1 hour each day)^38–41,43 and high SB/UB (screen time and TV >2 hours/day),^38–41,43 and sleep time (<10 hours/day)⁴³ were positively associated with Ow/Ob.

In the study by Balantekin et al.,⁴² the greatest percentage of girls with Ow/Ob (38%) were found in the cluster “Extreme Dieters,” with these being more likely to report attempting to reduce weight, dieting (eliminating snacking, sweets, and junk foods and reducing calories and the amount of food), increased exercise and unhealthy behaviors (use of laxatives/enemas, diuretics, diet pills, or appetite suppressants, smoking cigarettes, or self-induced vomiting).

In two studies, unhealthy lifestyles were positively associated with Ow/Ob, as well as alcohol use^44,45 and tobacco,⁴⁵ marijuana use,⁴⁵ more SB (>2 hours/day watching TV^44,45 and surfing the internet⁴⁵) and unhealthy eating (inadequate meals and food habits⁴⁵ and the consumption fast food and snacks⁴⁴).

Risk of Bias in Individual Studies

All studies presented high quality based on the tool JBI-MAStARI (Supplementary Appendix Table SA4).

Discussion

The results of this review indicate that the risk for Ow/Ob in children and adolescents can be influenced by the grouping of obesogenic behaviors (diet, PA, and SB/UB) and that the clusters presented in the studies, however distinct and complex, demonstrated the occurrence of healthy and unhealthy behaviors.

Ow/Ob is defined as the imbalance between energy intake and energy expenditure and apparently arises from the complex interaction of hereditary, social, socioeconomic, and, mainly, behavioral factors.^47,48 Although genetic factors have an influence on body weight, it is believed that behavioral factors, such as dietary factors, decreased PA, and increased sedentary leisure time, affect the early development of Ow/Ob.^48,49

Therefore, it was decided to synthesize the studies that analyzed these modifiable lifestyle risk factors according to more important and modifiable determinants for the control of obesity and chronic diseases⁵⁰ (diet, PA, and SB/UB) using LCA, since this statistic is used to identify subgroups of children and adolescents with different patterns of health-related risk behaviors² and the risk of developing Ow/Ob.

LCA is preferred in studies that include heterogeneous populations because it can identify the source of heterogeneity⁵¹ and it allows the joint analysis of many variables,² thus allowing it to identify similar patterns of responses, classifying homogeneous subgroups with similar characteristics of behaviors.⁵²

Different studies may give the same name to a dietary pattern, although this does not mean that the variates observed in these patterns were the same. There are possible variations due to the heterogeneity and specificity of the population (such as geographic, economic, and cultural characteristics), although several similarities could be observed.⁵³

In this review, the studies in the table demonstrated that independent of the name given to the clusters, the main characteristics of obesogenic risk behaviors were low weekly consumption of fruit and vegetables,^38–41,43 high consumption of fatty food, sugary snack foods, sweets, chips and fries,^38–40 low PA (<1 hour each day),^38–41,43 high SB [screen time and TV (2+ hours/day)],^38–40,43 and reduced sleep time (<10 hours/day).⁴³ These were positively associated with Ow/Ob.

In relation to diet, the results of this review are consistent with the worldwide trend toward Ow/Ob risk, with reports of a higher frequency of consuming obesogenic foods (rich in lipids, carbohydrates, bakery products, of animal origin, soft drinks, and sweets) and reduced consumption of fruits and vegetables, as highlighted by the WHO.⁵⁴

Studies carried out in several countries show the same trend of increasing consumption of potentially obesogenic foods (rich in sugar, such as breads, cakes, and cookies) by children and adolescents, for example, in the USA,^28,55 Brazil,⁵⁶ Mexico,⁵⁷ and European countries (regions in Italy, Estonia, Cyprus, Belgium, Sweden, Hungary, Germany, and Spain).^15,58

Food choices associated with increased Ow/Ob risk may be affected by a number of factors. A study conducted in the USA by French et al.⁵⁹ and another in China, Indonesia, Malaysia, and South Korea by Kelly et al.⁶⁰ indicate that television contributes to higher energy intake or increased dietary fat because food advertising is predominately related to fast food or convenience foods, most often sugary drinks, with low advertising rates for vegetables and fruits.

Cairns et al.⁶¹ concluded that the predominant advertising presence for food and beverages of low nutritional value for children occurs in both developed and low-income countries. Skidmore et al.⁶² emphasized that food choice is often affected by the distance and type of food establishments available, and Wang et al.⁶³ report that living in areas of low socioeconomic status leads to less access to supermarkets, resulting in the consumption of less healthy food.

In relation to PA, the results of this review showed that, in addition to an obesogenic diet, the prevalence of obesity was higher in the groups that presented PA (<1 hour each day) lower than the levels recommended by the WHO in 2019,⁶⁴ which recommends the practice of 60 minutes daily of MVPA for children and adolescents. PA level is likely determined by a complex mixture of biological, social, cultural, and environmental factors.⁶⁵

PA has several long-term benefits, with Janz et al.⁶⁶ and Baxter-Jones et al.⁶⁷ reporting that the practice of PA in childhood promoted positive effects on bone mineral density in adulthood. However, especially when performed with moderate to vigorous intensity, PA is favorably associated with multiple health indicators,⁶⁸ a greater effect on reducing cardiac risk and decreasing adiposity through BMI.⁶⁹ Dorling et al.⁷⁰ indicated that PA presented greater efficacy when combined with dietary interventions, as exercise has effects on the regulation of appetite.

Noonan et al.⁷¹ highlighted in their study that a way to encourage the regular practice of PA would be through school activities and social activities oriented around the family, as parents can help promote these healthy behaviors. DeWeese et al.²⁸ reported that, in addition to access to public and private recreational environments, such as parks or gyms for leisure activities, urbanization should also be effective in facilitating active transportation for individuals and thus affect eating and PA behaviors.

Considering SB, the results of this review showed that, in addition to an obesogenic diet and a low level of PA, the prevalence of obesity was higher in groups that presented SB [screen time—such as computer use, video games, and viewing other electronic media and TV (2+ hours/day)]. SB can be conceptualized as behavior that is distinct from any level of PA, with mild and moderate intensities that include sitting, lying down, watching television, and other forms of entertainment based on screen time⁷² and characterized by an energy expenditure of ≤1.5 METS in a seated or reclined posture.⁷³

Currently, children and adolescents have easy access to SB in their leisure time, especially TV viewing and electronic games, with more time spent in SB leading to less time spent in PA.⁷⁴ Wu et al.⁷⁵ concluded in their review that a higher frequency of PA or less SB equates to better health-related quality of life and a lower percentage of Ow/Ob.⁷⁶

SB may be dependent on culture, socioeconomic status, and/or lifestyle. Cui et al.⁷⁷ showed that Chinese children whose parents set rules for watching TV were less likely to be TV users. In contrast, it was also shown that easy access to TV, such as a TV in the bedroom, internet at home or in cyber cafes, or often watching TV with parents, were all associated with a higher probability of presenting SB (≥2 hours/day).⁷⁷ Janssen et al.⁷⁸ concluded that, from 7 to 15 years old, this SB tends to increase with age, so that by 15 years of age, typical sedentary time exceeded 75% of the waking day.

Current studies in several countries show the same tendency for SB. For example, regarding Japanese children, Ishii et al.⁷⁴ highlighted that on weekdays and weekends, children spent the most time watching TV, followed by online games. The study of Velde et al.⁷⁹ showed the same tendency in nine European countries (Austria, Belgium, Denmark, Iceland, the Netherlands, Norway, Portugal, Spain, and Sweden). The study from Biddle et al.⁸⁰ showed the same for Scottish adolescents, for whom television occupied the greatest amount of leisure time.

To reduce screen time and promote interactive games, the AAP⁸¹ recommends (1) limiting screen time for children <2 years of age; (2) limiting the total amount of entertainment screen time to 1–2 hours a day; (3) not watching TV during meals; (4) not putting a TV in the bedroom; (5) daily interactive play for children, ideally with parental involvement; and (6) monitoring the media that children are using and accessing, including any sites they visit and social media sites that they may be using.

The results of this review showed that there are some UBs, such as the use of laxatives/enemas, diuretics, self-induced vomiting, diet pills, appetite suppressants, smoking cigarettes, binge drinking, and excessive sleep time, which are associated with the prevalence of Ow/Ob.

Over the previous 5 years, new technologies and easy access to them have contributed to a rapid increase in average screen-time exposure for children and adolescents.⁸² Studies highlight that the longer children and adolescents spend in screen time (TV/videogame time) daily, the greater the sleep problems they may develop, such as, shorter sleep duration, with consequences on working memory and emotional control,⁸³ sleep efficiency, and verbal cognitive performance,⁸⁴ as well as behavioral health problems.⁸²

In this review, the studies of Pereira et al.⁴³ and Magee et al.³⁹ associated the risk of Ow/Ob with sleep time (<10 hours/day). These data are consistent with systematic reviews that point to a positive association between screen time (TV/video game time) and sleep problems,^85,86 with one of the consequences being sleep deprivation leading to fatigue, tiredness, and a greater propensity for SB and less PA.⁸⁷

Studies also show the association of cigarette smoking, and alcohol use, with the increased risk of obesity by adolescents,⁸⁸ alcohol, tobacco, and other drug use, with increased BMI⁸⁹ and, in the study by Liu et al.,⁹⁰ alcohol use was significantly associated with increased BMI in adolescents, although they found no significant association between illicit drug use and BMI.

The ideal is the joint identification of several modifiable risk factors, comparing them with other risk factors in terms of possible health benefits, and taking into account the specific context of each country and age group.⁹¹ The promotion of lifestyle modifications, including behavioral treatments, diet modification, and PA, can be the basis of primary/secondary prevention/treatment of pediatric obesity.^64,92

Strengths and Limitations of Review

One of the main strengths of this review is its analysis of studies that used the grouping of diet, PA, SB/UB, and their associations with Ow/Ob in children and adolescents. From an extensive search in different databases and with previously defined inclusion criteria, potential studies were identified; however, as these studies were heterogeneous, it was not possible to perform a meta-analysis and, therefore, the results were narratively described.

Studies have explored risk factors for Ow/Ob alone (e.g., diet and PA)^93,94; therefore, another strength of this review was the analysis considering the fact that obesogenic behaviors are not unique, with this joint analysis leading to a better understanding of which risk behaviors are the greatest predictors of Ow/Ob. This provides support so that future interventions can be adapted to these specific groups, considering behavioral differences.

Another aspect that sets this review apart is that some studies also examined other health behaviors, such as smoking or alcohol use and, particularly, sleep. Therefore, this review can provide important insights for the direction of public health initiatives aimed at combating obesity.

This review has several limitations. First, it is well known that the use of a cross-sectional design does not achieve the confidence level that meta-analyses of randomized clinical trials reach. Second, the studies did not take into account the differences of age and gender in the analysis of obesogenic behaviors. Finally, the studies included in this review were heterogeneous.

Conclusions: Implications for Future Research

This systematic review indicates that children and adolescents who adhere to dietary patterns composed of obesogenic foods are more likely to develop obesity, especially associated with low PA and excessive SB (especially screen time). It is suggested that to improve the understanding of modifiable lifestyle risk factors, future studies using LCA should use CA separately for age, gender, and environmental patterns, as this would provide more complete information for future interventions. Longitudinal studies could also be carried out, accompanying these children and adolescents to comprehend the changes in these obesogenic behaviors.

Footnotes

Authors' Contributions

All authors were involved in analyzing the studies, reviewing and interpreting the results, and writing the article.

Acknowledgment

We are grateful to the Postdoctoral National Program-Coordenação de Aperfeiçoamento de Pessoal de Nível Superior Coordination for the Improvement of Higher Education Personnel CAPES (PNPD/CAPES) (Programa Nacional de Pós Doutorado, PNPD-CAPES) that granted Rafaela Liberali a postdoctoral scholarship.

Funding Information

No funding was received for this article.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

References

Al-Hazzaa

, Abahussain

, Al-Sobayel

, et al. Lifestyle factors associated with overweight and obesity among Saudi adolescents. BMC Public Health, 2012; 12:354.

Patnode

, Lytle

, Erickson

, et al. Physical activity and sedentary activity patterns among children and adolescents: A latent class analysis approach. J Phys Act Health, 2011; 8:457–467.

WHO. Childhood overweight and obesity on the rise. 2016. Available at www.who.int/dietphysicalactivity/childhood/en/ Last accessed March 20, 2018.

Guo

, Wu

, Chumlea

, Roche

. Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence. Am J Clin Nutr, 2002; 76:653–658.

Katzmarzyk

, Barlow

, Bouchard

, et al. An evolving scientific basis for the prevention and treatment of pediatric obesity. Int J Obes, 2014; 38:887–905.

Umer

, Kelley

, Cottrell

, et al. Childhood obesity and adult cardiovascular disease risk factors: A systematic review with meta-analysis. BMC Public Health, 2017; 17:683.

Carnell

, Benson

, Chang

, et al. Neural correlates of familial obesity risk and overweight in adolescence. NeuroImage, 2017; 159:236–247.

Estrada

, Eneli

, Hampl

, et al. Children's hospital association consensus statements for comorbidities of childhood obesity. Child Obes, 2014; 10:304–317.

Sigmund

, Badura

, Sigmundová

, et al. Trends and correlates of overweight/obesity in Czech adolescents in relation to family socioeconomic status over a 12-year study period (2002–2014). BMC Public Health, 2018; 18:122.

10.

Landsberg

, Plachta-Danielzik

, Lange

, et al. Clustering of lifestyle factors and association with overweight in adolescents of the Kiel Obesity Prevention Study. Public Health Nutr, 2010; 13(1):1708–1715.

11.

Ottevaere

, Huybrechts

, Benser

, et al. Clustering patterns of physical activity, sedentary and dietary behavior among European adolescents: The HELENA study. BMC Public Health, 2011; 11:328.

12.

You

, Choo

. Adolescent overweight and obesity: Links to socioeconomic status and fruit and vegetable intakes. Int J Environ Res Public Health, 2016; 13:307.

13.

Ghanbari

, Mohammadpoorasl

, Jahangiry

, et al. Subgroups of lifestyle patterns among hypertension patients: A latent-class analysis. BMC Med Res Methodol, 2018; 18:127.

14.

Croezen

, Visscher

TLS

, Ter Bogt

NCW

, et al. Skipping breakfast, alcohol consumption and physical inactivity as risk factors for overweight and obesity in adolescents: Results of the E-MOVO project. Eur J Clin Nutr, 2009; 63:405.

15.

Aranceta

, Perez-Rodrigo

, Ribas

, Serra-Majem

. Sociodemographic and lifestyle determinants of food patterns in Spanish children and adolescents: The enKid study. Eur J Clin Nutr, 2003; 57(S1):S40.

16.

Wang

, Sun

, Zhai

, et al. The prevalence of depression and anxiety symptoms among overweight/obese and non-overweight/non-obese children/adolescents in China: A systematic review and meta-analysis. Int J Environ Res Public Health, 2019; 16:340.

17.

Berge

, Wall

, Bauer

, Neumark-Sztainer

. Parenting characteristics in the home environment and adolescent overweight: A latent class analysis. Obesity, 2010; 18:818–825.

18.

van Vliet

, Gustafsson

, Duchen

, Nelson

. Social inequality and age-specific gender differences in overweight and perception of overweight among Swedish children and adolescents: A cross-sectional study. BMC Public Health, 2015; 15:628.

19.

Bruening

, MacLehose

, Eisenberg

, et al. Friends like me: Associations in overweight/obese status among adolescent friends by race/ethnicity, sex, and friendship type. Child Obes, 2015; 11:722–730.

20.

Pitt

, Cameron

, Thornton

, et al. Dietary patterns of Australian children at three and five years of age and their changes over time: A latent class and latent transition analysis. Appetite, 2018; 129:207–216.

21.

Porcu

, Giambona

. Introduction to latent class analysis with applications. J Early Adolesc, 2017; 37:129–158.

22.

Funderburk

, Maisto

, Sugarman

, Wade

. The covariation of multiple risk factors in primary care: A latent class analysis. J Behav Med, 2008; 31:525–535.

23.

Laska

, Pasch

, Lust

, et al. Latent class analysis of lifestyle characteristics and health risk behaviors among college youth. Prev Sci, 2009; 10:376–386.

24.

Lanza

, Rhoades

. Latent class analysis: An alternative perspective on subgroup analysis in prevention and treatment. Prev Sci, 2013; 14:157–168.

25.

Smith

, Stein

, Cheng

, et al. Identifying high risk subgroups of MSM: A latent class analysis using two samples. BMC Infect Dis, 2019; 19:213.

26.

Breslau

, Reboussin

, Anthony

, Storr

. The structure of posttraumatic stress disorder: Latent class analysis in 2 community samples. Arch Gen Psychiatry, 2005; 62:1343–1351.

27.

Logan

, Hall

, Karch

. Suicide categories by patterns of known risk factors: A latent class analysis. Arch Gen Psychiatry, 2011; 68:935–941.

28.

DeWeese

, Ohri-Vachaspati

, Adams

, et al. Patterns of food and physical activity environments related to children's food and activity behaviors: A latent class analysis. Health Place, 2018; 49:19–29.

29.

Riggs

, Huh

, Chou

, et al. Executive function and latent classes of childhood obesity risk. J Behav Med, 2012; 35:642–650.

30.

Jeong

, Imboden

, Hansen

, et al. Heterogeneity of obesity-asthma association disentangled by latent class analysis, the SAPALDIA cohort. Respir Med, 2017; 125:25–32.

31.

Sacco

, Bucholz

, Spitznagel

. Alcohol use among older adults in the National Epidemiologic Survey on Alcohol and Related Conditions: A latent class analysis. J Stud Alcohol Drugs, 2009; 70:829–838.

32.

Dowd

, Keough

, Jakobson

, et al. A latent class analysis of young adult gamblers from the Manitoba Longitudinal Survey of Young Adults. Int Gambl Stud, 2019; 19:148–166.

33.

Moher

, Shamseer

, Clarke

, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev, 2015; 4:1.

34.

Rosato

, Baer

. Latent class analysis: A method for capturing heterogeneity. Soc Work Res, 2012; 36:61–69.

35.

Martin

, Booth

, Laird

, et al. Physical activity, diet and other behavioural interventions for improving cognition and school achievement in children and adolescents with obesity or overweight. Cochrane Database Syst Rev, 2018; 3:CD009728.

36.

Joanna Briggs Institute Critical Appraisal tools for use in JBI Systematic Reviews. Checklist for Analytical Cross Sectional Studies. Available at http://joannabriggs.org/research/critical-appraisal-tools.html Last accessed November 06, 2017.

37.

Joanna Briggs Institute. Joanna Briggs Institute reviewers' manual:. 2014. edition The Systematic Review of Prevalence and Incidence Data. The University of Adelaide: The Joanna Briggs Institute, 2014. Available at https://joannabriggs.org/assets/docs/sumari/ReviewersManual-2014.pdf Last accessed November 06, 2017.

38.

Huh

, Riggs

, Spruijt-Metz

, et al. Identifying patterns of eating and physical activity in children: A latent class analysis of obesity risk. Obesity, 2011; 19:652–658.

39.

Magee

, Caputi

, Iverson

. Patterns of health behaviours predict obesity in Australian children. J Paediatr Child Health, 2013; 49:291–296.

40.

Iannotti

, Wang

. Patterns of physical activity, sedentary behavior, and diet in

U.S

. Adolescents. J. Adolesc Health, 2013; 53:280–286.

41.

Wang

, Kim

, Su

, et al. Joint influence of individual choices, parenting practices, and physician advice on adolescent obesity, Nebraska, 2008. Prev Chronic Dis, 2014; 11:E175.

42.

Balantekin

, Birch

, Savage

. Patterns of weight-control behavior among 15 year old girls. Int J Eat Disord, 2015; 48:589–600.

43.

Pereira

, Katzmarzyk

, Gomes

. et al. Profiling physical activity, diet, screen and sleep habits in Portuguese children. Nutrients, 2015; 7:4345–4362.

44.

Laxer

, Brownson

, Dubin

, et al. Clustering of risk-related modifiable behaviours and their association with overweight and obesity among a large sample of youth in the COMPASS study. BMC Public Health, 2017; 17:102.

45.

Tabacchi

, Faigenbaum

, Jemni

, et al. Profiles of physical fitness risk behaviours in school adolescents from the ASSO project: A latent class analysis. Int J Environ Res Public Health, 2018; 15:1933.

46.

Lobo

, Assis

MAA

, Leal

, et al. Empirically derived dietary patterns through latent profile analysis among Brazilian children and adolescents from Southern Brazil, 2013–2015. PLoS One, 2019; 14:e0210425.

47.

Schneider

, Dumith

, Lopes

, et al. How do tracking and changes in dietary pattern during adolescence relate to the amount of body fat in early adulthood?. PLoS One, 2016; 11:e0149299.

48.

Lajunen

, Keski-Rahkonen

, Pulkkinen

, et al. Leisure activity patterns and their associations with overweight: A prospective study among adolescents. J Adolesc, 2009; 32:1089–1103.

49.

Dhana

, Haines

, Liu

, et al. Association between maternal adherence to healthy lifestyle practices and risk of obesity in offspring: Results from two prospective cohort studies of mother-child pairs in the United States. BMJ, 2018; 362:k2486.

50.

Brug

, van der Ploeg

, Loyen

, et al. DEDIPAC consortium. Determinants of diet and physical activity (DEDIPAC): A summary of findings. Int J Behav Nutr Phys Act, 2017; 14:150.

51.

Yang

, Shaftel

, Glasnapp

, Poggio

. Qualitative or quantitative differences? Latent class analysis of mathematical ability for special education students. J Spec Educ, 2005; 38:194–207.

52.

Nielsen

, Kent

, Hestbaek

, et al. Identifying subgroups of patients using latent class analysis: Should we use a single-stage or a two-stage approach? A methodological study using a cohort of patients with low back pain. BMC Musculoskelet Disord, 2017; 18:57.

53.

Oellingrath

, Svendsen

, Brantsæter

. Eating patterns and overweight in 9-to 10-year-old children in Telemark County, Norway: A cross-sectional study. Eur J Clin Nutr, 2010; 64:1272.

54.

World Health Organization (WHO). Diet, nutrition and the prevention of chronic diseases. World Health Organ Tech Rep Ser, 2003; 916:i–149.

55.

Ritchie

, Spector

, Stevens

, et al. Dietary patterns in adolescence are related to adiposity in young adulthood in black and white females. J Nutr, 2007; 137:399–406.

56.

Santos

NHAD

, Fiaccone

, Barreto

, et al. Association between eating patterns and body mass index in a sample of children and adolescents in Northeastern Brazil. Cad Saude Publica, 2014; 30:2235–2245.

57.

Romero-Polvo

, Denova-Gutiérrez

, Rivera-Paredez

, et al. Association between dietary patterns and insulin resistance in Mexican children and adolescents. Ann Nutr Metab, 2012; 61:142–150.

58.

Fernández-Alvira

, Bammann

, Eiben

, et al. Prospective associations between dietary patterns and body composition changes in European children: The IDEFICS study. Public Health Nutr, 2017; 20:3257–3265.

59.

French

, Story

, Jeffery

. Environmental influences on eating and physical activity. Ann Rev Public Health, 2001; 22:309–335.

60.

Kelly

, Hebden

, King

, et al. Children's exposure to food advertising on free-to-air television: An Asia-Pacific perspective. Health Promot Int, 2014; 31:144–152.

61.

Cairns

, Angus

, Hastings

, Caraher

. Systematic reviews of the evidence on the nature, extent and effects of food marketing to children. A retrospective summary. Appetite, 2013; 62:209–215.

62.

Skidmore

, Welch

, van Sluijs

, et al. Impact of neighbourhood food environment on food consumption in children aged 9–10 years in the UK SPEEDY (sport, physical activity and eating behaviour: Environmental determinants in young people) study. Public Health Nutr, 2010; 13:1022–1030.

63.

Wang

, Kim

, Gonzalez

, et al. Socioeconomic and food-related physical characteristics of the neighbourhood environment are associated with body mass index. J Epidemiol Commun Health, 2007; 61:491–498.

64.

World Health Organization. Guidelines on physical activity, sedentary behaviour and sleep for children under 5 years of age: Web annex: Evidence profiles. World Health Organization 2019. Available at www.who.int/iris/handle/10665/311663 (Last acceessed November 2019).

65.

Hallal

, Dumith

, Ekelund

, et al. Infancy and childhood growth and physical activity in adolescence: Prospective birth cohort study from Brazil. Int J Behav Nutr Phys Act, 2012; 9:82.

66.

Janz

, Letuchy

, Gilmore

JME

, et al. Early physical activity provides sustained bone health benefits later in childhood. Med Sci Sports Exerc, 2010; 42:1072–1078.

67.

Baxter-Jones

ADG

, Kontulaine

, Faulkner

, Bailey

. A longitudinal study of the relationship of physical activity to bone mineral accrual from adolescent to young adulthood. Bone, 2008; 43:1101–1107.

68.

Carson

, Lee

, Hewitt

, et al. Systematic review of the relationships between physical activity and health indicators in the early years (0–4 years). BMC Public Health, 2017; 17:985.

69.

Tarp

, Child

, White

, et al. International Children's Accelerometry Database (ICAD) Collaborators. Physical activity intensity, bout-duration, and cardiometabolic risk markers in children and adolescents. Int J Obes (Lond), 2018; 42:1639–1650.

70.

Dorling

, Broom

, Burns

, et al. Acute and chronic effects of exercise on appetite, energy intake, and appetite-related hormones: The modulating effect of adiposity, sex, and habitual physical activity. Nutrients, 2018; 22;10:1140.

71.

Noonan

, Boddy

, Fairclough

, Knowles

. Write, draw, show, and tell: A child-centred dual methodology to explore perceptions of out-of-school physical activity. BMC Public Health, 2016; 16:326.

72.

Boyington

, Joseph

, Fielding

, Pate

. Sedentary behavior research priorities—NHLBI/NIA sedentary behavior workshop summary. Med Sci Sports Exerc, 2015; 47:1291–1294.

73.

Gibbs

, Hergenroeder

, Katzmarzyk

, et al. Definition, measurement, and health risks associated with sedentary behavior. Med Sci Sports Exerc, 2015; 47:1295–1300.

74.

Ishii

, Shibata

, Adachi

, et al. School grade and sex differences in domain-specific sedentary behaviors among Japanese elementary school children: A cross-sectional study. BMC Public Health, 2017; 17:318.

75.

, Han

, Zhang

, et al. The influence of physical activity, sedentary behavior on health-related quality of life among the general population of children and adolescents: A systematic review. PLoS One, 2017; 12:e0187668.

76.

Goncalves

WSF

, Byrne

, Viana

, Trost

. Parental influences on screen time and weight status among preschool children from Brazil: A cross-sectional study. Int J Behav Nutr Phys Act, 2019; 16:27.

77.

Cui

, Hardy

, Dibley

, Bauman

. Temporal trends and recent correlates in sedentary behaviours in Chinese children. Int J Behav Nutr Phys Act, 2011; 8:93.

78.

Janssen

, Mann

, Basterfield

, et al. Development of sedentary behavior across childhood and adolescence: Longitudinal analysis of the Gateshead Millennium Study. Int J Behav Nutr Phys Act, 2016; 13:88.

79.

te Velde

, Bourdeaudhuij

, Thorsdottir

, et al. Patterns in sedentary and exercise behaviors and associations with overweight in 9-14-year-old boys and girls—A cross-sectional study. BMC Public Health, 2007; 7:16.

80.

Biddle

, Gorely

, Marshall

, Cameron

. The prevalence of sedentary behavior and physical activity in leisure time: A study of Scottish adolescents using ecological momentary assessment. Prev Med, 2009; 48:151–155.

81.

American Academy of Pediatrics. Children, Adolescents, and the Media. Council on communications and media. Pediatrics, 2013; 132:958–963.

82.

Parent

, Sanders

, Forehand

. Youth screen time and behavioral health problems: The role of sleep duration and disturbances. J Dev Behav Pediatr, 2016; 37:277–284.

83.

Warren

, Riggs

, Pentz

. Executive function mediates prospective relationships between sleep duration and sedentary behavior in children. Prev Med, 2016; 91:82–88.

84.

Dworak

, Schierl

, Bruns

, Strüder

. Impact of singular excessive computer game and television exposure on sleep patterns and memory performance of school-aged children. Pediatrics, 2007; 120:978–985.

85.

Cain

, Gradisar

. Electronic media use and sleep in school-aged children and adolescents: A review. Sleep Med, 2010; 11:735–742.

86.

Costigan

, Barnett

, Plotnikoff

, Lubans

. The health indicators associated with screen-based sedentary behavior among adolescent girls: A systematic review. J Adolesc Health, 2013; 52:382–392.

87.

Must

, Parisi

. Sedentary behavior and sleep: Paradoxical effects in association with childhood obesity. Int J Obes (Lond), 2009; Suppl 1(Suppl 1):S82–S86.

88.

Huang

, Lanza

, Anglin

. Association between adolescent substance use and obesity in young adulthood: A group-based dual trajectory analysis. Addict Behav, 2013; 38:2653–2660.

89.

Pasch

, Nelson

, Lytle

, et al. Adoption of risk-related factors through early adolescence: Associations with weight status and implications for causal mechanisms. J Adolesc Health, 2008; 43:387–393.

90.

Liu

, Yen

, Ko

, et al. Associations between substance use and body mass index: Moderating effects of sociodemographic characteristics among Taiwanese adolescents. Kaohsiung J Med Sci, 2010; 26:281–289.

91.

Lhachimi

, Nusselder

, Smit

, et al. Potential health gains and health losses in eleven EU countries attainable through feasible prevalences of the life-style related risk factors alcohol, BMI, and smoking: A quantitative health impact assessment. BMC Public Health, 2016; 16:734.

92.

McGovern

, Johnson

, Paulo

, et al. Treatment of pediatric obesity: A systematic review and meta-analysis of randomized trials. J Clin Endocrinol Metab, 2008; 93:4600–4605.

93.

Gazan

, Béchaux

, Crépet

, et al. Dietary patterns in the French adult population: A study from the second French national cross-sectional dietary survey (INCA2) (2006–2007). Br J Nutr, 2016; 116:300–315.

94.

Pérez-Rodrigo

, Gil Á, González-Gross

, et al. Clustering of dietary patterns, lifestyles, and overweight among Spanish children and adolescents in the ANIBES study. Nutrients, 2016; 8:11.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB