Relationship between Physical Activity and Diet among African-American Girls

Transcription

1 Relationship between Physical Activity and Diet among African-American Girls Russell Jago,* Tom Baranowski,* Sunmi Yoo, Karen W. Cullen,* Issa Zakeri,* Kathy Watson,* John H. Himes, Charlotte Pratt, Wanjie Sun, Leslie A. Pruitt,// and Donna M. Matheson// *Children s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, Texas; Department of Family Medicine, Inje University Sanggye Paik Hospital, Republic of South Korea; Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, Minnesota; Division of Epidemiology and Clinical Applications, National Heart, Lung and Blood Institute, Bethesda, Maryland; George Washington University, Biostatistic Center, Washington, DC; and Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, California. Address correspondence to Russell Jago, Children s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, 1100 Bates Street, Houston, TX rjago@bcm.tmc.edu Copyright 2004 NAASO Abstract JAGO, RUSSELL, TOM BARANOWSKI, SUNMI YOO, KAREN W. CULLEN, ISSA ZAKERI, KATHY WATSON, JOHN H. HIMES, CHARLOTTE PRATT, WANJIE SUN, LESLIE A. PRUITT, AND DONNA M. MATHESON. Relationship between physical activity and diet among African-American girls. Obes Res. 2004;12: 55S 63S. Objective: To examine the cross-sectional relationships between physical activity and dietary behaviors among 8- to 10-year-old African-American girls. Research Methods and Procedures: Two hundred ten 8- to 10-year-old African-American girls from four field centers participated. Computer Science and Applications (CSA) activity monitors were worn for 3 days. CSA data were expressed as mean CSA counts per minute, mean minutes of moderate to vigorous activity per day, and mean metabolic equivalents (METS) per minute. Two nonconsecutive 24- hour dietary recalls were analyzed for kilocalories; percent kilocalories from fat; daily servings of fruit, 100% fruit juice, and vegetables; sweetened beverages; and water consumption. Height and weight were measured, and information on household income, material possessions, and participant age were obtained. Results: All three expressions of physical activity were significantly negatively associated with percentage calories from fat (r to 0.177, p 0.01), and mean METS per minute were significantly positively associated with percentage calories from carbohydrate (r 0.149, p 0.05) after controlling for household income, material possessions, field center, and total caloric intake. Income was inversely associated with percentage calories from fat. Discussion: Physical activity and dietary fat consumption were inversely related among African-American girls. Efforts to prevent obesity in preadolescent African-American girls should focus on increasing physical activity and lowering dietary fat consumption. Key words: energy balance, physical activity, dietary fat Introduction The prevalence of overweight among all 6 to 11 year olds in the U.S. increased from 11.3% in 1988 to 1994 to 15.3% in 1999 to 2000 (1). Among African-American boys and girls, the prevalence of overweight increased 10 percentage points over the same period (1). Childhood obesity increases the chances of adult obesity (2) and the risk of cardiovascular disease (3). Obesity is an imbalance between energy consumed and energy expended. Physical activity is the most malleable component of energy expenditure (4), and physical activity declines throughout childhood (5) and adolescence (6). To maintain energy balance, lower levels of physical activity should be accompanied by lower caloric intake (7). Dietary behaviors that have been associated with lower BMI include the consumption of fruit (8,9) and vegetables (9). In contrast, high dietary fat intake and increased consumption of sweetened beverages are associated with increased caloric intake and body weight (10,11). Several studies have shown an association between physical activity and the dietary intake of adults and adolescents. Active adults consumed more calories than sedentary adults (12) and ate healthier diets, consuming more fiber, less saturated fat, fewer servings of fried food and sweets, and more servings of fruit and vegetables (12 14). Adults who ate breakfast regularly had more adequate micronutrient intake and fewer calories from fat than non-breakfast eaters (15), and breakfast eaters were more physically active than non-breakfast eaters (15). Adolescent athletes consumed more calories, more dairy products, cereals, fruit, fruit 55S

2 juices, and salads than less active adolescents (16). Active male and female adolescents were more likely to consume breakfast than their inactive counterparts (17). These analyses, however, did not control for total caloric intake. There is little information about the relationships between physical activity and dietary behaviors among preadolescent children in general, and African Americans in particular. Because there is a high prevalence of obesity among African-American girls (18), information is needed about energy balance and the relationship between physical activity and dietary intake in this group. The goal of the Girls Health Enrichment Multi Site Center Studies (GEMS), 1 a multicenter research program sponsored by the National Heart, Lung, and Blood Institute, was to develop and test separate interventions to prevent excessive weight gain in preadolescent African-American girls. Four field centers (Baylor College of Medicine, Houston, TX; University of Minnesota, Minneapolis, MN; University of Memphis, Memphis, TN; and Stanford University, Palo Alto, CA) independently developed and pilot-tested their own interventions but shared some common evaluation measurements. Each field center randomly assigned 20 to 40 girls per treatment arm to either an active intervention or comparison group for 12 weeks. Common baseline measurements included 3-day assessment of physical activity using an accelerometer and dietary intake using two 24-hour dietary recalls. Using baseline data, we examined whether physical activity was positively associated with percentage calories consumed from carbohydrate, fruit, 100% fruit juice, and vegetable (FJV) intake, numbers of meals and snacks eaten, and breakfast consumption, and negatively associated with sweetened beverage intake and percentage calories from fat among 8- to 10-year-old African-American girls. 1 Nonstandard abbreviations: GEMS, Girls Health Enrichment Multi Site Center Studies; FJV, fruit, 100% fruit juice, and vegetable; CSA, Computer Science and Applications; METS, metabolic equivalents. Research Methods and Procedures Subjects In the winter and spring of 2001, the four field centers enrolled a total of to 10-year-old African-American girls who completed baseline measures. Data are presented for the 198 subjects from whom complete data were obtained. Inclusion criteria were BMI 25% for age and sex at the Minnesota and Memphis field centers, BMI 50% for age and sex at the Baylor field center, and BMI 50% for age and sex or at least one parent with a BMI 25 kg/m 2 at the Stanford field center. Girls were excluded from the study if they were taking medications or had medical illnesses affecting growth, had conditions limiting participation in physical activity, or had conditions limiting participation in measurements. Parents provided written informed consent, and girls gave assent to participate in the pilot study. The Human Subjects Review Board at each participating institution approved the study protocol. Measures Subject Characteristics. Demographic characteristics including girl s age, parent education, total household income, and material possessions (an index of the number of material possessions, such as televisions, video cassette recorders, and cell phones in the home) (19), were obtained by parental self-report. Girls height and weight were measured according to the GEMS study protocol (20). Height was measured twice using a stadiometer (Shorr Height Measuring Board, Olney, MD), and the mean of the two readings was calculated. Body weight was measured twice using a calibrated scale (Seca 770 model scale; Vogel and Halke, Hamburg, Germany), and the mean of the two readings was calculated. BMI (kilograms per meter squared) was computed. Physical Activity Measurement. Physical activity levels were recorded using the Computer Science Application (CSA) accelerometer (Computer Science Applications, Shalimar, FL). (Since this study was conducted, the CSA has been rebranded as the MTI Actigraph, but because the models used in this study were purchased as CSA monitors, they will be referred to as such throughout this paper.) The CSA monitor has been shown to be a reliable and valid measure of physical activity level in children (21,22). To reduce the number of times that the monitor was removed and the associated risks of forgetting to put the monitor back on, girls were instructed to wear the CSA monitor continuously, including during sleep, for 3 days, except while showering or swimming. Girls were asked to record in a log the time when the CSA monitor was taken off. All participants wore the monitor above the right hip. After 3 complete days, CSA monitors and logs were collected from the girls. More detail on the use of CSA activity monitors in the GEMS study can be found elsewhere (23). Participant CSA data were included if the girl wore the instrument for at least 1 of 3 days for 800 minutes or more between 6:00 AM and 12:00 PM during the baseline assessment. Counts recorded between these times were used for all analyses because they were considered to be representative of normal waking hours. CSA data were expressed in three ways. First, the number of counts recorded over the 3-day period was divided by the number of minutes worn to provide mean daily counts per minute, an expression of the total volume of activity. Second, CSA counts were converted to mean daily minutes of moderate to vigorous intensity physical activity using the 1952 counts/min cut-point (24) to provide an indication of the amount of time spent engaged in health-enhancing physical activity. Third, CSA counts were expressed as mean metabolic equivalents 56S

3 (METS) per minute, using an age-dependent formula (27), to provide an estimate of the intensity of the physical activity in which the girls engaged. The equation for this conversion is as follows: ( counts/min) ( age [years]) ( counts/min age [years]) (27). It should be noted that, although the criteria used to classify CSA counts as moderate to vigorous activity and convert CSA counts to METs were originally developed for adults (24), both have been extensively used for the analysis of data from children (25 28). These measures were used for analysis to enable the interpretation of the separate relationships between volume, amount of time, and intensity of activity and dietary variables. Dietary Assessment. Each girl completed two nonconsecutive 24-hour dietary recalls. Field center staff obtained the information using a laptop computer and Minnesota Nutrient Data System revised software (NDS-R 4.02). NDS-R 4.02 features a multiple-pass approach that prompts for complete food descriptions and preparation methods and includes 16,000 foods and values for 117 nutrients and nutrient ratios (29). The recalls were analyzed for average daily intake of kilocalories, percent kilocalories from fat and from carbohydrates, and daily servings of FJV, sweetened beverages, and water. The number of meals and snacks eaten and whether breakfast was eaten per day were also ascertained. More detail on dietary measurement in GEMS can be found elsewhere (30). Data Analysis. Descriptive statistics including means, ranges, SD, and scatter plots were calculated for each variable. Values 50% of the range of preceding values that also visually appeared out of range from histograms were considered outliers and were removed from the dataset. Normality was assessed using the Kline criteria for skewness ( 3.0) and kurtosis ( 10.0) (31). Variables were logtransformed to approximate normality if the normality assumption was violated. Bivariate Pearson and Spearman correlations were assessed between each of the physical activity measures and each of the dietary variables, but because the results from both correlations were highly similar, only Pearson correlations are reported. Because no significant relationships were found between physical activity and FJV consumption when calculated separately (likely because of a high frequency of zero consumption), these three variables were combined to obtain servings of FJV. Each dietary variable was separately regressed on the three expressions of physical activity using hierarchical linear regression (with backward elimination). To test for differences between the four field centers, the models incorporated field centers as a fixed effect. The models also included covariates to adjust for total household income, caloric intake, material possessions, and the interactions between field center and each covariate. All nonsignificant terms (BMI, age, and parental education) were subsequently removed from the models in the backward deletion process, along with all of the interactions between field centers and covariates. Total caloric intake was included as a covariate to control for the relationship of total calories with physical activity. Field center was included as a fixed effect in the regression models. This approach was taken because, although using a random effect model is the usual approach, it has been reported that including field center as a random effect in studies that include a small number of field centers ( 10) can provide only minimal information about the group and is not necessary (32). Furthermore, the use of field center as a fixed effect was supported by the intraclass correlations for the three expressions of physical activity, percentage calories from fat, and percentage calories from carbohydrates, the three chief outcome variables, which indicated that the design effects were small enough to ignore. The above models were selected as the primary method of analysis because they allowed examination of the primary question of interest, the relationship of dietary intake variables to physical activity, controlling for the relation of total calories with physical activity. In addition we also controlled for income, material possessions, and field center. Results Descriptive statistics for the 198 subjects who met the CSA inclusion criteria are shown in Table 1. The average time that the CSA monitors were worn was 17.3 h/d (96% of the total time between 6:00 AM and 12:00 PM). The mean age of the sample was 9.2 years, and the mean BMI was 22.2 kg/m 2, indicating that the majority of the girls were at risk of being overweight (BMI 85%) or were, in fact, overweight (BMI 95%) (1). The mean daily consumption of servings of FJV was less than two servings, and the percentage calories consumed from fat exceeded 34%, indicating that, on average, these girls were not meeting the U.S. dietary guidelines (33). The mean daily water consumption was also less than one serving. Bivariate Pearson correlations between mean daily CSA counts per minute, mean daily minutes of moderate to vigorous intensity physical activity, and mean daily METS per minute from 6:00 AM to 12:00 PM, and all of the demographic and dietary variables are shown in Table 2. A significant negative correlation (r 0.16, p 0.05) was obtained between physical activity and age when mean daily counts per minute were used as the indicator of physical activity, but this relationship was not evident for the number of minutes spent engaged in moderate to vigorous intensity physical activity. Because the calculation of daily METS is dependent on subject age and CSA counts, the correlation between age and METS is not interpretable. The significant negative correlations between mean daily counts per minute and BMI (r 0.15, p 0.05) and mean daily 57S

4 Table 1. Descriptive statistics for demographic characteristics, dietary measures,* and daily physical activity (assessed between 6:00 AM and 12:00 PM) among 8- to 10-year-old African-American girls at Memphis, Minnesota, Stanford, and Baylor field centers Variable N Range Median Mean SD Demographic variables Age (years) to BMI (kg/m 2 ) to Highest education achieved to Total household income to Material possessions (no. of items) to Dietary variables Caloric intake (kcal/day) to Percent calories from fat per day to Percent calories from carbohydrates per day to Fruit, juice and vegetable (svgs/day) to Sweetened beverages (svgs/day) to Water (svg/day) to Number of meals per day to Number of snacks per day to Ate breakfast both days (yes 1) to Physical activity variables Mean daily CSA counts per minute to Sum of minutes of moderate PA to Average METS per minute to Minutes CSA worn (6:00 AM to 12:00 PM) to * Dietary measures (including number of meals and snacks) are means assessed from 2 days of recall. Categorical variable where 1 6th grade, 2 8th grade, 3 some high school, 4 high school grad or GED, 5 technical school, 6 some college, 7 college graduate, and 8 postgraduate study. Categorical variable where 1 $5000, 2 $5000 to $9999, 3 $10,000 to $19,999, 4 $20,000 to $29,999, etc., and 12 $100,000. Minutes of moderate to vigorous PA are minutes when the CSA count is 1952/min. METS per minute are calculated using Trost et al. s (2002) formula: ( counts/min) [ age (years)] counts/min age (years). PA, physical activity. minutes of moderate to vigorous intensity physical activity and BMI (r 0.20, p 0.01) indicated that active girls, in general, and more moderately vigorous active girls in particular, had lower BMI than less active girls. Significant negative correlations were evident among all three expressions of physical activity and percentage calories from fat (r 0.19 to 0.24, p 0.01). Significant positive correlations were evident among all three expressions of physical activity and percentage calories from carbohydrate (r 0.18 to 0.24, p 0.05). The results of hierarchical linear regression analyses with mean daily CSA counts per minute as the independent variable and percentage calories from fat and carbohydrate as dependent variables are shown in Table 3. The first model relates the mean daily CSA counts per minute to the dependent variable after controlling for household income, material possession, field center, and total caloric intake. There was a significant negative relationship between mean daily CSA counts per minute and percentage calories from fat after controlling for the other variables. The relationship between mean daily CSA counts per minute and percentage calories from carbohydrate was nonsignificant after controlling for the other variables. Results of hierarchical linear regression analyses with minutes of moderate to vigorous physical activity as the independent variable and percentage calories from fat and carbohydrate as dependent variables were virtually identical to those with mean daily CSA counts per minute [R 2 of 58S

5 Table 2. Pearson correlation coefficients between physical activity as measured by mean daily CSA counts per minute, mean daily minutes of moderate to vigorous physical activity (MVPA),* and mean METs per minute with demographic characteristics, dietary measures, and BMI Variable N Mean daily CSA counts per minute Mean daily minutes of MVPA Mean METS per minute Demographic variables Age (years) ** BMI (kg/m 2 ) Highest Education (years) Total Household Income Material possessions (no. of items) Dietary variables Caloric Intake (kcal/day) Percent calories from fat per day Percent calories from carbohydrates per day Fruit, juice, and vegetable (svgs/day) Sweetened beverages (svgs/day) Water (svg/day) Number of meals per day Number of snacks per day Ate breakfast both days (yes 1) * Minutes of moderate to vigorous physical activity are assessed as minutes when CSA counts are 1952/min. METs are calculated using Trost et al. s (27) formula: ( counts/min) [ age (years)] counts/min age (years) Categorical variables where 1 6th grade, 2 8th grade, 3 some high school, 4 high school grad or GED, 5 technical school, 6 some college, 7 college graduate, and 8 Postgraduate study. Categorical variable where 1 $5000, 2 $5000 to $9999, 3 $10,000 to $19,999, 4 $20,000 to $29,999, etc., and 12 $100,000. Correlation is significant at p ** Correlation not interpretable. Correlation is significant at p for percentage calories from fat (p 0.04) and for percentage from carbohydrate (p 0.11)]. Results of hierarchical linear regression analyses with mean METS per minute as the independent variable and percentage calories from fat and carbohydrate as dependent variables are also shown in Table 3. While results were similar to those for mean daily CSA counts per minute, the positive relationship between METS and percentage calories from carbohydrate remained significant after controlling for all terms. A similar degree of variance for percentage calories from fat and from carbohydrate was explained by mean daily CSA counts and by mean METs per minute. Hierarchical linear regression analyses with minutes of moderate to vigorous activity, mean daily CSA counts, and mean METs per minute as the independent variables and FJV, meals, snacks, breakfast consumption, and sweetened beverage consumption as the dependent variables in five separate models revealed no significant relationships (data not shown). Discussion Physical Activity and Dietary Intake All three physical activity measures (i.e., CSA counts per minute, minutes of moderate to vigorous physical activity, and mean METS per minute) were negatively associated with the percentage calories consumed from fat after controlling for household income, material possessions, field center, and total caloric consumption. This negative relationship has been reported before (34), but this is the first study to obtain this relationship after controlling for total caloric intake among African-American girls. Thus, rather than simply showing that more active people eat more in 59S

6 Table 3. Hierarchical linear regression analyses (backward elimination) of mean daily CSA counts per minute between 6:00 AM and 12:00 PM and mean METS* per minute on the percent calories from fat and carbohydrate (independently) Dependent variable Independent variables Std coefficients p Percent calories from fat Mean daily CSA counts per minute Field center: reference (Memphis) Minnesota Stanford Baylor Total household income Material possessions Caloric intake R Percent calories from carbohydrate Mean daily CSA counts per minute Field center: reference (Memphis) Minnesota Stanford Baylor Total household income Material possessions Caloric intake R Percent calories from fat Mean METS per minute Field center: reference (Memphis) Minnesota Stanford Baylor Total household income Material possessions Caloric intake R Percent calories from carbohydrate Mean METS per minute Field center: Reference (Memphis) Minnesota Stanford Baylor Total household income Material possessions Caloric intake R * METs are calculated using Trost et al. s formula: ( counts/min) [ age (years)] counts/min age (years). Controlling for caloric intake, field center, total household income, and material possessions. (BMI and age backward deleted). general, and thereby, eat more fat, this study suggests that active African-American girls tend to consume diets lower in fat than less active girls. The obtained relationships might be stronger if more days of assessment were included for both the 24-hour dietary recalls and the CSA physical activity monitoring. There are three possible interpretations of this finding. First, more health-conscious girls might be more active and select healthier diets. Second, perhaps girls 60S

7 who consume greater quantities of fat feel sluggish because foods high in fat have the potential to slow gastric emptying, or girls who consume less fat may choose to engage in increased amounts of physical activity. Third, less active girls could spend more time watching television, and particularly television commercials, which contain a large number of advertisements for high-fat snack products (35). Children s requests and parental purchases of snack products have been associated with the frequency with which products have been advertised on television in other studies (36). Children who ate meals while watching television consumed less vegetables and increased amounts of snack foods (37). Physical activity levels, when converted to METS (but not mean CSA counts or minutes of moderate to vigorous physical activity), were positively associated with the percentage calories consumed from carbohydrate, even after controlling for household income, material possessions, caloric intake, and field center differences. There is normally little variation in protein consumption (38), and lower fat consumption is often associated with higher carbohydrate intake (39). As such, the positive association between METS and percentage calories consumed from carbohydrate is consistent with the observed relationship between higher activity and decreased dietary fat intake. It is not clear why this relationship was only observed when METS were used as the indicator of physical activity. It could be that there is a closer association between the intensity of physical activity in which girls engage and their carbohydrate intake than with either the volume of activity or minutes of moderate to vigorous physical activity. The relationship could also be an artifact of the MET conversion formula, and therefore, further research that examines these possibilities is warranted. Activity Levels and BMI Physical activity was negatively correlated with BMI, indicating that active girls had lower adiposity levels than sedentary girls. Physical activity increases lean body mass and reduces adiposity (40), and longitudinal analyses have shown that regular physical activity aids weight control by maintaining energy expenditure and preventing weight gain (41,42). Conversely the lower levels of activity could be the result of increased adiposity. Further research is required to assess these two possibilities. Energy Balance Equation Contrary to predictions from the energy balance equation (43), physical activity was not significantly associated with total caloric intake in correlation analyses. This finding could be at least partially due to the limitations of both the physical activity and dietary measurements used in this study. In a study by Trost et al. (44), 5 days of objective monitored physical activity were required to achieve a reliability of 0.8 in children, whereas we only had three. Furthermore, the two 24-hour dietary recalls used in this study also may not have been sufficient to provide a reliable assessment of diet (45,46). Expressing CSA Data Of the three expressions of physical activity, all of which were significantly and negatively correlated with percentage of calories from fat, only the mean METS per minute was not significantly correlated with BMI. This lack of association may be because of an artifact of the MET conversion formula, which was originally designed for adults. Because this formula has been used with children (26), and there is a lack of a validated alternative among children, a MET formula should be developed for children. Analyses should be conducted to examine the relationships between dietary intakes and both the volume and intensity of physical activity in children. Strengths and Limitations To the best of our knowledge, this is the first study to examine the relationships between physical activity and dietary behaviors of African-American girls using an objective measure of physical activity and controlling for total caloric intake. The limitations of this study include the limited reliability of dietary assessment (only 2 days) (45,46) and of physical activity (only 3 days) (44), with the resulting attenuation of correlations and regression estimates. Furthermore, the dietary data were measured by self-report, which suffers from several known biases, including underreporting (47,48). As previously mentioned, although extensively used for research with children, the criteria used to classify CSA counts as moderate to vigorous activity and to convert CSA counts to METs were originally developed for adults (24). The cross-sectional design of the study precluded establishing causal relationships among the variables of interest. Although statistically significant associations were found between physical activity and percentage calories from fat, all of the models accounted for 15% of the variance, suggesting that factors other than those recorded in this study were influencing the relationship between the dietary and physical activity patterns of African-American girls in this study. Furthermore, because field center was included as a fixed effect in the regression models, the results hold only for the four field centers studied in this investigation and can not be generalized to the population of all field centers. However, the study identified significant relationships that require further investigation. In conclusion, low levels of physical activity and increased dietary fat consumption were positively related among African-American girls. Future research should ex- 61S

8 amine the causal directionality of this relationship with additional possible explanatory factors. Acknowledgments This research was largely funded by National Heart, Lung, and Blood Institute Grants U01 HL65160, U01 HL62662, U01 HL62663, U01 HL62668, and U01 HL This work is a publication of the United States Department of Agriculture (USDA/ARS) Children s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, and had been funded, in part, with federal funds from the USDA/ARS under Cooperative Agreement The contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement from the US government. References 1. Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence and trends in overweight among US children and adolescents, JAMA. 2002;288: Whitaker RC, Wright JA, Pepe MS, Seidel KD. Predicting obesity in young adulthood from childhood and parental obesity. N Engl J Med. 1997;337: Freedman DS, Dietz WH, Srinivasan SR, Berenson GS. The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart study. Pediatrics. 1999;103: Jakicic JM, Clark K, Coleman E, et al. Appropriate intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc. 2001;33: Sallis JF, Alcaraz JE, McKenzie TL, Hovell MF. Predictors of change in children s physical activity over 20 months variations by gender and level of adiposity. Am J Prev Med. 1999;16: Kimm SY, Glynn NW, Kriska A, et al. Decline in physical activity in black girls and white girls during adolescence. N Engl J Med. 2002;347: Esparza J, Fox C, Harper IT, et al. Daily energy expenditure in Mexican and USA Pima Indians: low physical activity as a possible cause of obesity. Int J Obes Relat Metab Disord. 2000;54: Crooks DL. Food consumption, activity, and overweight among elementary school children in an Appalachian Kentucky community. Am J Phys Anthropol. 2000;112: Maskarinec G, Novotny R, Tasaki K. Dietary patterns are associated with body mass index in multiethnic women. J Nutr. 2000;130: Ludwig DS, Peterson KE, Gortmaker SL. Relation between consumption of sugar sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet. 2001;357: Harnack L, Stang J, Story M. Soft drink consumption among US children and adolescents: nutritional consequences. J Am Dietet Assoc. 1999;99: Brodney S, McPherson RS, Carpenter RA, Welten D, Blair SN. Nutrient intake of physically fit and unfit men and women. Med Sci Sports Exerc. 2001;33: Eaton CB, McPhillips JB, Gans KM, et al. Cross-sectional relationship between diet and physical activity in two southeastern New England communities. Am J Prev Med. 1995;11: Matthews CE, Hebert JR, Ockene IS, Saperia G, Merriam PA. Relationship between leisure-time physical activity and selected dietary variables in the Worcester Area Trial for counseling in hyperlipidemia. Med Sci Sports Exerc. 1997;29: Barnard JA, Tapsell LC, Davies PSW, Brenninger VL, Storlien LH. Relationship of high energy expenditure and variation in dietary intake with reporting accuracy on 7 day food records and diet histories in a group of healthy adult volunteers. Eur J Clin Nutr. 2002;56: Cavadini C, Decarli B, Grin J, Narring F, Michaud PA. Food habits and sport activity during adolescence: differences between athletic and non-athletic teenagers in Switzerland. Eur J Clin Nutr. 2000;54:S Aarnio M, Winter T, Kujala U, Kapiro J. Associations of health related behavior, social relationships, and health status with persistent physical activity and inactivity: a study of Finnish adolescent twins. Br J Sports Med. 2002;36: Kimm SY, Barton BA, Obarzanek E, et al. Obesity development during adolescence in a biracial cohort: the NHLBI Growth and Health study. Pediatrics. 2002;110:e Baranowski T, Baranowski J, Cullen KW, et al. 5 a Day achievement badge for African-American boy scouts: pilot outcome results. Prev Med. 2002;34: Gordon CC, Chumlea WC, Roche AF. Stature, recumbent length, and weight. In: Lohman TG, Roche AF, Martorell R, eds. Anthropometric Standardization Reference Manual. Champaign, IL: Human Kinetics; 1988, pp Janz KF. Validation of the CSA accelerometer for assessing children s physical activity. Med Sci Sports Exerc. 1994;26: Trost SG, Ward DS, Moorehead SM, Watson PD, Riner W, Burke J. Validity of the Computer Science and Applications (CSA) activity monitor in children. Med Sci Sports Exerc. 1998;30: Treuth M, Sherwood NE, Baranowski T, et al. Physical activity self-report and accelerometry measures from the Girls Health Enrichment Multi-site Studies (GEMS). Prev Med. 2004;38:S Freedson PS, Melanson E, Sirad J. Calibration of the Computer Science and Application accelerometer. Med Sci Sports Exerc. 1998;30: Ekelund U, Sjöström M, Yngve A, et al. Physical activity assessed by activity monitor and doubly labeled water. Med Sci Sports Exerc. 2001;33: Trost S, Kerr LM, Ward DS, Pate R. Physical activity and determinants of physical activity in obese and non-obese children. Int J Obes Relat Metab Disord. 2001;25: Trost SG, Pate RR, Sallis JF, et al. Age and gender differences in objectively measured physical activity in youth. Med Sci Sports Exerc. 2002;34: Pate RP, Freedson PS, Sallis JF, et al. Compliance with 62S

9 physical activity guidelines: prevalence in a population of children and youth. Ann Epidemiol. 2002;12: Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient database. J Am Dietet Assoc. 1988;88: Cullen KW, Himes JH, Baranowski T, et al. Validity and reliability of a behavior-based food coding system for measuring fruit, 100% fruit juice, vegetable, and sweetened beverage consumption: results from the Girls Health Enrichment Multi-Site Studies (GEMS). Prev Med. 2004;38:S Kline RB. Principles and Practice of Structural Equation Modeling. New York, NY: Guilford Press; Snijders T, Bosker R. Multilevel Analyses: An Introduction to Basic and Advanced Multilevel Modeling. Thousand Oaks, CA: Sage; USDA. Nutrition and Your Health: Dietary Guidelines for Americans, 5th ed. Washington, DC: United States Department of Health and Human Services; Mattison I, Wirfalt E, Gullberg B, Berglund G. Fat intake is more strongly associated with lifestyle factors than with socio-economic characteristics, regardless of energy adjustment approach. Eur J Clin Nutr. 2001;55: Jeffrey DB, McLellarn RW, Fox DT. The development of children s eating habits: the role of television commercials. Health Educ Qtly. 1982;9: Taras HL, Sallis J, Patterson TL, Nader PR, Nelson JA. Television s influence on children s diet and physical activity. J Develop Behav Pediatr. 1989;10: Coon KA, Goldberg J, Rogers BL, Tucker KL. Relationships between use of television during meals and children s food consumption patterns. Pediatrics. 2001;107: Putnam J, Gerroir S. Trends in the U.S. food supply, In: Frazao E, ed. Agriculture Information Bulletin. Washington, DC: USDA; 1999, pp Nicklas T, Elkasabany A, Srinivasan SR, Berenson GS. Trends in nutrient intake of 10-year-old children over two decades. Am J Epidemiol. 2001;153: Rippe JM, Hess S. The role of physical activity in the prevention and management of obesity. J Am Dietet Assoc. 1998;98:S Williamson DF, Madans J, Anda RF, Kleinman JC, Kahn HS, Byers T. Recreational physical activity and ten-year weight change in a US national cohort. Int J Obes Relat Metab Disord. 1993;17: Sherwood NE, Jeffrey RW, French SA, Hannan PJ, Murray DM. Predictors of weight gain in the Pound of Prevention study. Int J Obes Relat Metab Disord. 2000;24: Lahti-Koski M, Pietinen P, Heliovaara M, Vartiainen E. Associations of body mass index and obesity with physical activity, food choices, alcohol intake, and smoking in the FINRISK studies. Am J Clin Nutr. 2002;75: Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. Using objective physical activity measures with youth: how many days are needed. Med Sci Sports Exerc. 2000;32: Brekke MJ, Brekke ML, Quiter ES, Peters JR, Mullis RM, Hunninghake DB. Appropriate days of measuring intake of dietary fat and cholesterol. Ann Nutr Metab. 1992;36: Larken FA, Metzner HL, Guire KE. Comparison of three consecutive and three random day records of dietary intake. J Am Dietet Assoc. 1991;91: Schoeller DA. Limitations in the assessment of dietary energy intake by self-report. Metabolism. 1995;44: Lissner L, Heitmann BL, Bengtsson C. Population studies of diet and obesity. Br J Nutr. 2000;83:S S