Table 1 Characteristics of studies included in the systematic review (n = 17)
First author (publication year) | Country | Original project | Sample size (girls %) | Age group (mean age) | Method used to derive clusters | Number of clusters | Correlates associated with clusters |
|---|---|---|---|---|---|---|---|
De Bourdeaudhuij (2013) [34] | European countriesa | ENERGY | 766 (52.9%) | 10 – 12 years (11.5 girls /11.7 boys) | Two step cluster analysis (hierarchical and non-hierarchical methods) | Boys 4 Girls 4 | BMI and waist circumference |
Gorely (2007) [15] | UKb | STIL | 1,371 (62.0%) | Mean 14.7 years (sd = 0.92, range 12.5–17.6 years) | Cluster analysis (Ward’s method and k-means) | Boys 5 Girls 5 | None |
Huang (2016) [17] | Chinac | Not reported | 951 (50.5%) | 9 – 13 years (11.0) | Cluster analysis (hierarchical) | Boys 5 Girls 5 | Sociodemographic factors and sports team participation |
Kim (2016) [26] | USA | YRBS | 12,081 (49.4%) | 9th – 12th grades (adolescents) | Latent class analysis | Boys 4 Girls 4 | Obesity |
Lazarou (2009) [35] | Cyprus | CYKIDS | 1,140 (53.4%) | 10 – 13 years (10.7) | Principal component analysis | 8 | None |
Marshall (2002) [36] | USA and UK | Not reported | USA: 1,750 (59.0%) UK: 744 (85.0%) | USA: mean 12.9 years (sd = 0.92) UK: mean 13.0 years (sd = 0.94) | Cluster analysis | Boys 3 Girls 3 | Age, nationality, ethnicity, and BMI |
Melkevik (2011) [25] | Norway | HBSC | 4,848 (48.0%) | 13, 15, and 16 years | Latent profile analysis | Boys 6 Girls 6 | Overweight |
Nelson (2005) [37] | USA | Add Health | 1,1957 (50.0%) | Mean age (wave I) 14.9 years (sd = 0.12) | Cluster analysis* | 7 | Meet PA guidelines |
Nelson (2006) [23] | USA | Add Health | 1,1957 (50.0%) | Mean age (wave II) 15.8 years (sd = 11.6) | Cluster analysis* | 7 | Health risk behaviors and other weekly activities, and self-esteem |
O'Neill (2017) [38] | Ireland | GUI | 8,568 (48.9%) | 9 – 13 years | Two step cluster analysis | Boys 4 Girls (no coherent cluster type found) | Weight status |
Patnode (2011) [24] | USA | IDEA and ECHO | 720 (51.1%) | Mean age 14.7 years (sd = 1.8) | Latent class analysis | Boys 3 Girls 3 | Grade, race, parent education, live with 2 parents, overweight, weight status, free or reduced-price lunch |
Ramos (2012) [39] | Spain | HBSC | 21,811 (53.1%) | 11 – 18 years | Cluster analysis (general linear models) | 3 | Biopsychosocial health |
Spengler (2015) [16] | Germany | MoMo | 2,083 (50.5%) | 11 – 13 years 14 – 17 years | Cluster K-means | Boys 8 Girls 7 | Age and socioeconomic status |
Taverno Ross (2016) [40] | USA | TRACK | 495 (55.4%) | 5th (baseline) and 7th grades (children) | Latent class analysis | Boys 3 Girls 3 | Socio-demographics, Individual-level factors and Interpersonal-level factors; School-level factors# |
te Velde (2007) [13] | European countriesd | CSS | 12,538 (50.1%) | 8.8 – 13.8 years (11.4) | Cluster K-means | Boys 5 Girls 5 | Overweight |
Wang (2006) [41] | Singapore | Not reported | 780 (61.8%) | 11 – 14 years | Cluster analysis (hierarchical methods) | Boys 3 Girls 3 | None |
Wang (2012) [42] | Singapore | Not reported | 847 (61.0%) | 10 – 16 years | Latent profile analysis | 5 | None |