Cohen 2014.

Methods	Study design: prospective cohort study. Analyses methods for cohorts: linear regression with participant‐level random‐effects model used to examine whether physical activity, diet and environmental exposures were associated prospectively with changes in bodyweight and % body fat. Only variables that were significant were combined into a single multivariate model. How were missing data handled? Only the participants who had valid data for all 3 assessment periods were analysed (n = 265 (87%) compared to n = 303 who were enrolled). Number of study contacts: 3 (baseline in grade 8, 2 follow‐up visits in tenth/eleventh grade or eleventh/twelfth grade). Period of follow‐up (total period of observation): 5 years. Periods of recruitment: 2007, as the follow‐up across grades 10‐12 occurred during 2009‐2011. Sample size justification adequately described? No. Study authors also mentioned that a limitation in the study was the relative small sample size. Sampling method: random sample. Control participants of the TAAG cohort from 2 sites (San Diego, Minneapolis) used (532 eligible girls). For present analysis, 303 girls were randomly selected from 7 different high schools in these sites. Study objective: to study correlates of physical activity and nutrition behaviours and change in BMI percentile and body fat among adolescent girls. Study population: 13‐ to 18‐year‐old girls at high schools in San Diego and Minneapolis.
Participants	Baseline characteristics (reported for 1 overall group) Age (mean in years): 13.9 (SD 0.4). Sex: 100% girls. Ethnicity: 54.3% non‐Hispanic white, 27.1% Hispanic, 4.2% black, 7.9% Asian, 6.4% other. Education: mother's education: 37.4% high school or lower, 59.6% college or higher, 3% unknown. Income: households in poverty: 5.5% (SD 3.6). Pubertal stage: NR. Parental BMI: NR. Child total energy: NR. Child total fat: NR. Child total protein: NR. Child total CHO: NR. Child physical activity: sedentary minutes/day: 533.3 (SD 61.5); moderate‐vigorous physical activity (minutes/day): 22.2 (SD 10.1); number of sports/physical activity teams/classes in past year: 3.4 (SD 3.3); currently taking physical education at school: 85.7%. Child physical inactivity or screen time or both: (Min/day): 212.4 (SD 116.6). Child CVD risk (excluding fatness): NR. Child body fatness: BMI: 22.1 (SD 5.2); overweight (BMI ≥ 85th percentile): 30.9%; obese (BMI ≥ 95th percentile): 15.1%; % body fat: 29.3 (SD NR). Included criteria: 8th grade girls who were control participants enrolled in the TAAG study cohort from 2 sites. Excluded criteria: NR. Brief description of participants: school girls, in grade 8 across 7 high schools from 2 sites in the USA (San Diego and Minneapolis/St Paul). During study period, participants were aged 13‐18 years. Total number completed in cohort study: 265 (87%). Total number enrolled in cohort study: 303.
Interventions	Description of exposure for cohorts Time span: 3 years. Dietary assessment method used: validated FFQ completed by participants. Frequency of assessments: single FFQ completed twice (9th or 10th grade) and (11th and 12th grade). No dietary assessment at baseline (8th grade). See Table 9; Table 10; Table 11; Table 12; Table 13; Table 14; Table 15; Table 16; Table 17; Table 18 for details of total fat intake exposure per outcome.
Outcomes	BMI BMI percentile (%). Body fat Body fat (%).
Identification	Sponsorship source: National Health, Lung and Blood Institute. Country: USA. Setting: high schools, San Diego and Minneapolis. Comments: NA. Author's name: Deborah A Cohen. Institution: RAND Corporation. Email: dcohen@rand.org. Declaration of interests: yes. "None of the authors have any financial relationships relevant to this article or other conflicts of interest to disclose." Study ID: Cohen 2014. Type of record: journal article. Trial ID: TAAG.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Were adequate outcome data for cohorts available? All outcomes	Low risk	Attrition low (13%; 38/303). Children with incomplete data did not differ from children with complete data in terms of ethnicity, mother's education and age (data NR).
Was there matching of less‐exposed and more‐exposed participants for prognostic factors associated with outcome or were relevant statistical adjustments done? All outcomes	High risk	Data analysis did not adjust for pubertal stage, parental BMI and total energy intake at baseline.
Did the exposures between groups differ in components other than only total fat? All outcomes	Low risk
Can we be confident in the assessment of outcomes? All outcomes	High risk	Methods used to measure body fat were inconsistent during the study (skinfold thickness measurements at baseline, BIA during follow‐up).
Can we be confident in the assessment of exposure? All outcomes	High risk	No baseline dietary assessment. Unclear whether they received any training or assistance regarding the completion of the FFQ during follow‐up.
Can we be confident in the assessment of presence or absence of prognostic factors? All outcomes	Low risk	Repeated measurements of physical activity data were performed (accelerometer data for 6 consecutive days). 16.8% of data imputed. Self‐report of variables such as age, ethnicity and mother's education was acceptable at this age.
Was selection of less‐exposed and more‐exposed groups from the same population? All outcomes	Low risk	All control participants of the TAAG cohort.