Television viewing and change in body fat from preschool to early adolescence: The Framingham Children s Study

(2003) 27, 827 833 & 2003 Nature Publishing Group All rights reserved 0307-0565/03 $25.00 www.nature.com/ijo PAPER Television viewing and change in body fat from preschool to early adolescence: The Framingham Children s Study MH Proctor 1 *, LL Moore 1, D Gao 1, LA Cupples 2, ML Bradlee 1, MY Hood 1 and RC Ellison 1 1 Section of Preventive Medicine and Epidemiology, Boston University School of Medicine, Boston, MA, USA; and 2 Department of Epidemiology and Biostatistics, Boston University School of Public Health, Boston, MA, USA OBJECTIVE: To prospectively examine the relation between television watching and body fat change in children from preschool to early adolescence. METHODS: In a longitudinal study, 106 children were enrolled during preschool years (mean age 4.0 y) and followed into early adolescence (mean age 11.1 y). Parents completed an annual questionnaire on the child s television and video habits. Body mass index (BMI), triceps skinfolds, and sum of five skinfolds were recorded yearly at annual clinic visits. Longitudinal statistical analyses were carried out using mixed modeling procedures to control for potential confounding by a number of factors. RESULTS: Television watching was an independent predictor of the change in the child s BMI, triceps, and sum of five skinfolds throughout childhood. Its effect was only slightly attenuated by controlling for the baseline body fat, level of physical activity (as measured repeatedly by Caltrac accelerometer), percent of calories from fat, total calorie intake, or the parents BMI or education. By age 11, children who watched 3.0 h or more of television per day had a mean sum of skinfolds of 106.2 mm, compared with a mean sum of skinfolds of 76.5 mm for those who watched less than 1.75 h per day (P ¼ 0.007). Furthermore, the adverse effect of television viewing was worse for those children who were also sedentary or had a higher-fat diet. CONCLUSIONS: Children who watched the most television during childhood had the greatest increase in body fat over time. Healthy lifestyle education designed to prevent obesity and its consequences should target television-watching habits of children. (2003) 27, 827 833. doi:10.1038/sj.ijo.0802294 Keywords: television watching; childhood obesity; physical activity Introduction The childhood shows the man As morning shows the day. John Milton, Paradise Regained Risk factors for cardiovascular disease (CVD) often have their origin in childhood, when many lifestyle behaviors become established. 1 Dietary patterns and physical activity levels that contribute to the development of obesity often *Correspondence: Dr MH Proctor, Section of Preventive Medicine and Epidemiology, Boston University School of Medicine, 715 Albany Street, Boston, MA 02118, USA. E-mail: mproctor@bu.edu Received 24 July 2002; revised 9 December 2002; accepted 26 January 2003 track into adulthood, particularly among children who are obese during adolescence. 2 4 Adult obesity is a strong risk factor for diabetes, hypertension, and ischemic heart disease. 5 Recent studies showing a dramatic rise in the rates of obesity in both children and adults in the US 6 7 are cause for alarm. An increasing number of hours spent watching television is one mechanism that may underlie these rapidly rising rates. Some 8 11 but not all studies 12,13 among children find a positive association between television viewing and obesity. Differing methodologies and conflicting results have made it difficult for investigators to reach a consensus on the importance of this link. The Framingham Children s Study (FCS) has shown that as the level of physical activity increases among preschool children, anthropometric measures of body fatness decrease. 14

828 In this report, we use data from the FCS to examine the relation between the amount of time spent watching television throughout childhood and the change in body fatness from preschool to the early adolescent years. We also explore the extent to which dietary fat and physical activity levels may impact on the association between television and body fat change. Materials and methods The FCS is a longitudinal study of early childhood determinants of diet and physical activity. The study began in 1987, when 106 preschool children and their parents were enrolled. The subjects were third- and fourth-generation offspring of the original Framingham Heart Study cohort, who were identified by means of questionnaires sent to surviving members of the original cohort as well as their offspring. One child per family, between the ages of 3 and 5 y, was enrolled. Details of the recruitment and enrollment have been previously published. 14 The parents and children have been evaluated annually at a clinic examination visit. The parents completed questionnaires at each annual clinic visit reporting the average amount of time that the child spent watching television each day. Since this was a longitudinal study, some changes to the assessment tools were necessary as the children aged. Once children were in school most of the day, the television questionnaire was modified to ascertain the average number of hours spent watching television on schooldays and weekend days during the school year, and on weekdays and weekends during the summer. Starting in about second grade, the children began also to spend some time with video games, so this was added to the questionnaire. We estimated the average hours of television and video per day during each year as a weighted average of the reported television and video hours during the school year and in the summer. Throughout the remainder of this report, we will refer collectively to television and video hours per day as television hours per day. We measured height, weight, and skinfold thickness for children and parents each year at the clinic exam. Weight (to 1 the nearest 4 pound) was measured using a standard counterbalance scale, and height was measured (to the nearest 1 4 in) using a measuring bar on the same scale. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. On each visit, the examiners used Lange calipers and a standard protocol to take two measures each (to the nearest millimeter) of triceps, subscapular, suprailiac, abdominal, and thigh skinfolds for each child. We also gathered information on the following potential confounders: the child s age, gender, baseline body fat, physical activity level, total energy intake, and percent of calories from fat as well as the mother s and father s age, education and BMI at baseline. Since television viewing may have its effects on the child s body fat change through its association with physical activity level or dietary habits, we also examined the effect of these factors in combination with television viewing on body fat change. Physical activity was assessed using an electronic motion sensor, the Caltrac Accelerometer, an instrument that has been shown to provide valid estimates of physical activity in children. 15 The children wore the device attached at the waist over the right hip for three to five consecutive days each year. During the first year, there were four separate 5- day monitoring periods, while in subsequent years, there were either one or two such periods. The Caltrac was programmed to provide unitless activity counts to give a measure of total movement. This information was combined with data on the amount of time the instrument was worn to estimate the average number of Caltrac counts per hour during each year of monitoring. Dietary intake was assessed in year 1 by means of four sets of 3-day diet records. In each subsequent year, one or two sets of 3-day records were collected. The study nutritionist taught each family how to complete the diaries, including how to estimate portion sizes. The same nutritionist debriefed the parents, children, and other caregivers as needed to clarify information on the diaries. The mean total energy intake and percent of calories from fat consumed each year were estimated by entering the food diary information into the Nutrition Data System of the University of Minnesota. 16 Data for the present analysis were available for 103 children. Statistical methods The primary goal of the analysis was to determine whether children who watched more television throughout childhood had greater increases in their BMI or subcutaneous skinfolds over time than children who watched less. We first calculated the child s mean television viewing hours per day during each of eight exam cycles and then ranked each child s hours per day each year. These ranks were used to separate the children into tertiles (low, middle, high) of television viewing each year. Thus, a given child might have been ranked in the low tertile during the first year, the middle tertile the next year, and the high tertile the following year. This enabled us to account for changing television viewing patterns over time. We used longitudinal data analysis methods to estimate the effect of television-viewing habits over time. We then constructed mixed regression models to estimate the association between the tertile of television watching and three outcome measures: BMI, triceps skinfold, and sum of all five skinfolds. The models produce estimates of adjusted least-squared means for BMI, triceps, and sum of skinfolds at each age according to the amount of television watched at the time. We then plotted the adjusted least-squared means to analyze trends in body fat change throughout childhood. We controlled potential confounders in the mixed regression models by adding the child s sex, baseline age (as a continuous variable), baseline anthropometry (BMI or skin-

folds as appropriate to the model), and each parent s level of education. We then added maternal and paternal BMI as surrogates for a possible genetic correlation in body fat. Finally, we added factors at each year that might either serve as confounders or causal intermediates in the relation between television watching and body fat change: physical activity (Caltrac counts per hour categorized as highest tertile vs lower two tertiles), total energy intake (kcal/day), and mean percent of calories from fat (o34% vs Z34%). Cutpoints for activity and percent of calories from fat were chosen on the basis of earlier analyses (manuscripts under review). For example, in our earlier work, children in the highest activity tertile had significantly lower gains in body fat throughout childhood while those in the low and middle tertiles had almost identical changes in body fat over 8 y. All statistical analyses were performed using SAS statistical analysis system. Results Baseline characteristics of children and parents according to tertile of television watched per day at baseline are shown in Table 1. At baseline, when they averaged 4.0 y of age, children in the lowest tertile watched an average of 1.0 h of television per day, while those in the middle and upper tertiles watched 1.6 and 2.4 h per day, respectively. At baseline, there was no apparent association between age, activity level, or anthropometry measurements with television-viewing habits. The children in the highest tertile of television viewing did have the highest total energy intake per day. However, there was no clear pattern of differences in the intakes of fat, carbohydrate, protein or micronutrients, although calcium intakes did seem a little higher among the children in higher tertiles of television watching per day. It is interesting to note that the children who watched the most television had parents with higher BMI levels. Parental education level was inversely correlated with child s television viewing time. In Table 2, we examine the differences in mean BMI at the end of follow-up (mean age ¼ 11.1 y) according to the mean hours of television watched per day throughout childhood. In this table, we have classified hours of television watched per day into three categories: o1.75 h, 1.75 to o3.0, and Z3.0 h/day. By the time of early adolescence, BMI was highest for those children who watched the most television during childhood and lowest for those who watched the least (mean BMI ¼ 20.9 vs 18.6 kg/m 2 for those watching Z3.0 h vs o1.75 h/day). The results were similar for all anthropometry indices. At the end of follow-up at 11 y, those who watched 3.0 h or more of television per day had a mean sum of five skinfolds of 106.2 mm compared with a mean of 87.6 mm for those watching between 1.75 and 3.0 h of 829 Table 1 Baseline characteristics of children and parents according to hours of television and video per day Tertile of television and video watched per day Lowest Middle Highest n ¼ 34 n ¼ 34 n ¼ 35 Mean7s.d. Child s baseline data Television (h/day) 1.170.5 1.670.7 2.471.6 Age (y) 3.970.7 4.170.8 4.170.8 Activity (caltrac counts/h) 10.771.8 10.672.2 10.071.6 BMI (kg/m 2 ) 16.371.1 16.171.2 16.271.1 Triceps skinfold (mm) 11.572.5 10.972.4 11.572.1 Sum of skinfolds (mm) 41.6710.8 41.5710.4 43.0711.0 Energy intake (kcal/day) 14867261 15497226 16017342 Percent of calories from fat 33.675.1 32.773.8 33.573.6 Percent of calories from carbohydrate 53.877.0 55.475.8 53.675.3 Percent of calories from protein 13.972.2 13.472.3 13.372.0 Calcium (mg)/1000 kcal 463.27152.6 486.5794.2 518.07121.9 Magnesium (mg)/1000 kcal 126.3728.4 126.2721.8 121.7717.1 Vitamin A (mcg RE)/1000 kcal 433.77214.6 427.67163.8 429.27192.8 Parents baseline data Mother s activity (caltrac counts/h) 5.971.6 6.072.0 5.972.6 Father s activity (caltrac counts/h) 6.671.9 6.171.8 6.271.8 Mother s BMI (kg/m 2 ) 23.774.4 24.774.3 25.074.5 Father s BMI (kg/m 2 ) 25.972.5 27.974.0 28.173.4 Percent of subjects Child s gender (% male) 50.0 61.8 71.4 Mother s education (some college or more) 73.5 58.8 54.3 Father s education (some college or more) 72.7 72.7 54.3

830 Table 2 Level of body fat at the end of follow-up according to hours of television and video per day Anthropometry at end of follow-up a Mean hours of TV and video per day throughout childhood Sample Body mass index (kg/m 2 ) Triceps skinfolds (mm) Sum of skinfolds (mm) mean7s.e. mean7s.e. mean7s.e. Adjusted for age, sex, and baseline anthropometry o1.75 h n ¼ 25 18.670.6 16.271.1 76.577.1 1.75 to o3.0 h n ¼ 49 19.670.4 17.170.8 87.675.0 3.0+ h n ¼ 20 20.970.7 20.371.2 106.278.1 P-value for linear trend b 0.0430 0.039 0.028 Adjusted for age, sex, baseline anthropometry and parents BMIs o1.75 h n ¼ 25 19.170.6 17.171.1 82.276.9 1.75 to o3.0 h n ¼ 49 19.570.4 17.170.8 86.874.8 3.0+ h n ¼ 20 20.970.6 20.271.1 106.077.5 P-value for linear trend b 0.109 0.059 0.052 Adjusted for age, sex, baseline anthropometry and physical activity o1.75 h n ¼ 25 18.770.6 16.471.1 77.977.2 1.75 to o3.0 h n ¼ 49 19.670.4 17.170.8 87.575.0 3.0+ h n ¼ 20 20.870.7 20.071.2 104.778.1 P-value for linear trend b 0.096 0.081 0.057 Adjusted for age, sex, baseline anthropometry, total energy and percent of calories from fat o 1.75 h n ¼ 25 18.770.6 16.371.1 77.677.0 1.75 to o3.0 h n ¼ 49 19.670.4 17.270.8 88.474.9 3.0+ h n ¼ 20 20.770.7 19.871.2 102.878.0 P-value for linear trend b 0.098 0.094 0.073 a Age range ¼ 10 12 y (mean age ¼ 11.1 y). b Pr0.05 for all anthropometry outcomes, comparing highest and lowest TV hours per day (o1.75 vs Z3.0 h). television per day, and 76.5 mm for those watching o1.75 h/ day (P-value for trend ¼ 0.028). We then controlled for parental body fat, physical activity, and dietary variables (percent of calories from fat and total energy intake). Controlling for these factors led to slight attenuation of the effect of television watching on body fat change. In all cases, those who watched three or more hours of television per day (group 3) had statistically significant higher BMI, triceps, and sum of five skinfolds than children watching less than 1.75 h/day (group 1). In Figures 1 3, we examine the change in the child s BMI, triceps skinfolds, and sum of five skinfolds according to the tertile of television viewing. Here, we use tertile of television watching to maximize the power of the analyses. In all of these figures, the anthropometry means are adjusted for the child s gender, baseline age, and anthropometry level; mean physical activity level, percent of calories from fat, and total energy intake per day at each age; and each parent s education level and BMI. Adjustment for the child s baseline anthropometry measurement was the only factor that had any effect on the shape of the curves at all. (The effect of this adjustment was to narrow the difference between the curves very slightly.) Children in the highest television-watching group showed the greatest increases in their mean BMI from ages 4 11 y (Figure 1). Those in the intermediate group of television hours per day had intermediate gains in BMI and those who BMI (kg/m 2 ) 1 22 20 18 16 14 Lowest Tertile Middle Tertile Highest Tertile Figure 1 Body mass index from 4 to 11 y according to tertile of television and video per day. 1 Adjusted for child s sex, baseline age and BMI, mean percent of calories from fat (o34 vs Z34%), mean total energy intake/day, mean physical activity (highest vs lower two tertiles), and parents education levels and BMIs). watched the least television had the smallest BMI gains. In Figures 2 and 3, the effects are similar. Those who watched the most television had the greatest increases in triceps and sum of skinfolds and those who watched the least television had the smallest gains in body fat. To explore further the extent to which physical activity or diet may explain the effect of television watching on body

Triceps Skinfold (mm) 1 20 18 16 14 12 10 Lowest Tertile Middle Tertile Highest Tertile Figure 2 Triceps skinfolds from 4 to 11 y according to tertile of television and video per day. 1 Adjusted for child s sex, baseline age and triceps, mean percent of calories from fat (34 vs Z34%), mean total energy intake per day, mean physical activity (highest vs lower two tertiles), and parents education levels and BMIs.) Sum of Skinfolds (mm) 1 100 90 80 70 60 50 40 30 Low activity/high TV Low activity/low TV High activity/hightv High activity/low TV Figure 4 Sum of five skinfolds from ages 4 to 11 y according to television watching and physical activity. 1 Adjusted for child s sex, baseline age and sum of five skinfolds, mean percent of calories from fat (o34 vs Z34%), mean energy intake/day, parents education levels and BMIs. 2 Low vs high activity ¼ lower two tertiles vs highest tertile of caltrac counts per hour. Low vs high TV ¼ mean TV and video hours per day: o2h vs Z2h.) 831 Sum of Skinfolds (mm) 1 100 80 60 40 20 Lowest Tertile Middle Tertile Highest Tertile Figure 3 Sum of skinfolds from 4 to 11 y according to tertile of television and video per day. 1 Adjusted for child s sex, baseline age and sum of five skinfolds, mean percent of calories from fat (34 vs Z34%), mean total energy intake/day, mean physical activity (highest vs lower two tertiles), and parents education levels and BMIs.) Sum of Skinfolds (mm) 1 110 100 90 80 70 60 50 40 30 High Fat/High TV High Fat/Low TV Low Fat/High TV Low Fat/Low TV 4 5 6 7 8 9 10 11 Figure 5 Sum of five skinfolds from ages 4 to 11 y according to television watching and percent of calories from fat. 1 Adjusted for child s sex, baseline age and sum of five skinfolds, mean percent of calories from fat (o34 vs Z34%), mean energy intake/day, parents education levels, and BMIs. 2 Low vs high fat ¼ mean percent of calories from fat, o34% vs Z34%. Low vs high TV ¼ mean TV and video hours per day: o2h vs Z2h.) fat change, we look at the combined effects of these factors in Figures 4 and 5. Figure 4 shows the combined effects of television and activity level. Those children who were the most sedentary (low activity) and who watched the most television (high TV) had the greatest gains in their sum of five skinfolds from ages 4 to 11 y. Conversely, children who were active (high activity) and who watched little television (low TV) had a sum of five skinfolds that was nearly 40 mm smaller by 11 y of age. Children who watched a lot of television but who were also very active gained slightly more body fat than active children who did not watch as much television. Finally, Figure 5 explores the combined effects of television watching and dietary fat intake. Clearly, those children who watched the most television and who had a high-fat diet (Z34% of calories) gained the most body fat over time, and those who watched the least television and had a lower-fat diet (o34% of calories) had the smallest gains in body fat. Discussion It is clear from the report from the Third National Health and Nutrition Examination Survey 7 that an increasing number of children in the US are overweight, and many are obese. These results from the FCS confirm that the number of hours spent watching television or playing video games is an

832 important risk factor for the development of excess body fat during childhood. This adverse effect is partly a consequence of the combined effects of television-viewing habits and a sedentary lifestyle or a high-fat diet. In this study, the children at greatest risk for having excess body fat by the time of early adolescence were those having low levels of physical activity or a high-fat diet in combination with watching more hours of television per day. Our results differ from those of Robinson et al, 12 who reported on a group of adolescent females followed for 24 months; they found only a weak association between television viewing and either obesity or physical activity. The weak effect in that study may reflect the older ages of the children or the shorter follow-up time. It is also possible that children in southern California (where year-round outdoor exercise is easier to achieve) who watch a lot of television may be less sedentary in general. DuRant et al 13 followed a group of 3 to 4-y-old children over 3 y and found no increase in measures of obesity (despite finding that children who watched more television were physically less active). In the FCS, however, we found that the differences in body fat associated with excess television viewing became more apparent over time. Thus, it may be that follow-up time in DuRant et al s study was insufficient to detect important changes in body fat associated with televisionviewing habits. In this study, we also noted that the children who watched the least television had parents with higher levels of education (and parents who had lower levels of body fatness themselves). These findings are consistent with other studies showing an inverse association between socioeconomic status and obesity (as well as the subsequent incidence of ischemic heart disease). 17 Since television-viewing habits are likely to be similar in parents and children, the children who watch the least television may have parents who are leaner as a consequence of their own lower levels of television watching. The FCS has a number of unique characteristics that strengthen these results. Most importantly, it is a prospective study with repeated measures of television viewing and anthropometry during childhood. These repeated assessments yield more stable estimates of the child s televisionviewing habits during childhood as well as long-term estimates of body fat change. Consequently, this study provides important new evidence that television watching is a risk factor for change in body fat, not simply reflective of more obese children tending to watch more television as a consequence of their obesity making it difficult to exercise. Despite this, some may still argue that children who gained the most body fat were simply predisposed to become fat (on the basis of genetic factors) and that these same children may have been predisposed to a sedentary lifestyle (and hence higher levels of television watching) on the basis of a similar genetic correspondence in activity. To address this, we controlled for body fat of both the mothers and fathers as well as the child s baseline body fat in the multivariable models; we found that the amount of television watched per day was still an independent predictor of body fat change. There are a number of possible mechanisms by which television viewing may have an adverse effect on body fat change during childhood. Children who watch more television and play more video games may be less active in general and may have less favorable dietary habits. After controlling for these factors, television watching still remained an independent predictor of body fat change in this study. However, further analyses suggested that there seems to be some interaction between these factors (diet, activity, and television habits), making it impossible to simply control for the effects of diet or activity in the multivariable models. In addition, at least one earlier study has also shown that television may have a hypometabolic effect; in that study, they found that resting energy expenditure while watching television (for both obese and normal weight children) was significantly lower than when they were doing nothing at all. 18 The relation between television-viewing habits and dietary factors is a complex issue that warrants examination in future studies. It may be that television viewing alters the child s intake of foods in ways that are not adequately reflected by total energy intake or percent of calories from fat. In this study, the correlation between hours of television watched per day and percent of calories from fat was 0.13. Children who watch a lot of television may have very different patterns of snacking and food consumption than children who watch less television. This might, for example, lead to alterations in glycemic index or fiber intake, thereby changing the child s subsequent risk of obesity. Alterations in snacking patterns may be a response to idleness or a consequence of the more direct influence of television advertising. Some authors report that weekly viewing hours correlated with children s caloric intake and reported food requests, and that food purchases made by their parents were influenced by television. 19 Borzekowski and Robinson 20 comment that their findings suggest that it takes only one or two exposures to a 10 30 s food commercial to influence preschoolers short-term preferences for specific food products. The results of this study highlight the need to target families of very young children (preschool and early elementary school) for interventions designed to reduce the risk of adolescent obesity and the subsequent risk of adult chronic diseases. Several recent studies have reported successful school-based and family-based interventions to reduce obesity through less television use and increased physical activity. 21 24 Those results, in combination with our own, suggest that healthy lifestyle education programs designed to prevent adult cardiovascular disease should include components targeting television watching habits of children.

Acknowledgements This study is supported by Grant no. HL35653 from The National Heart, Lung and Blood Institute. References 1 Berenson GS, Srinivasan SR, Hunter SM, Nicklas TA, Freedman DS, Shear CL, Webber LS. Risk factors in early life as predictors of adult heart disease: the Bogalusa Heart Study. Am J Med Sci 1989; 298: 141 151. 2 Kelder SH, Perry CL, Klepp KI, Lytle LL. Longitudinal tracking of adolescent smoking, physical activity, and food choice behaviors. Am J Public Health 1994: 84; 1121 1126. 3 Lauer RM, Lee J, Clarke WR. Factors affecting the relationship between childhood and adult cholesterol levels: the Muscatine Study. Pediatrics 1988; 82: 309 318. 4 Serdula MK, Ivery D, Coates RJ, Freedman DS, Williamson DF, Byers T. Do obese children become obese adults? A review of the literature. Prev Med 1993; 22: 167 177. 5 Hu FB, Leitzmann MF, Stampfer MJ, Colditz GF, Willett WC, Rimm EB. Physical activity and television watching in relation to risk for type 2 diabetes mellitus in men. Arch Intern Med 2001; 161: 1542 1548. 6 Kuczmarski RJ, Flegal KM, Campbell SM, Johnson CL. Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. JAMA 1994; 272: 205 211. 7 Troiano RP, Flegal KM, Kuczmarski RJ, Campbell SM, Johnson CL. Overweight prevalence and trends for children and adolescents: the National Health and Nutrition Examination Surveys, 1963 to 1991. Arch Pediatr Adolesc Med 1995; 149: 1085 1091. 8 Dietz WH, Gortmaker SL. Do we fatten our children at the television set? Obesity and television viewing in children and adolescents. Pediatrics 1985; 75: 807 812. 9 Wong ND, Hei TK, Qaqundah PY, Davidson DM, Bassin SL, Gold KV. Television viewing and pediatric hypercholesterolemia. Pediatrics 1992; 90: 75 79. 10 Jeffery RW, French SA. Epidemic obesity in the United States: are fast foods and television viewing contributing? Am J Public Health 1998; 88: 277 280. 11 Hernandez B, Gortmaker SL, Colditz GA, Peterson KE, Laird NM, Parra-Cabrera S. Association of obesity with physical activity, television programs and other forms of video viewing among children in Mexico City. Int J Obes Relat Metab Disord 1999; 23: 845 854. 12 Robinson TN, Hammer LD, Killen JD, Kraemer HC, Wilson DM, Hayward C, Taylor CB. Does television viewing increase obesity and reduce physical activity? Cross-sectional and longitudinal analyses among adolescent girls. Pediatrics 1993; 91: 273 280. 13 DuRant RH, Baranoski T, Johnson M, Thompson WO. The relationship among television watching, physical activity and body composition of young children. Pediatrics 1994; 94: 449 455. 14 Moore LL, Nguyen UDT, Rothman KJ, Cupples LA, Ellison RC. Preschool physical activity level and change in body fatness in young children: The Framingham Children s Study. Am J Epidemiol 1995; 142: 982 988. 15 Noland M, Danner F, Dewalt K, McFadden M, Kotchen JM. The measurement of physical activity in young children. Res Q Exerc Sport 1990; 61: 146 153. 16 Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient data base. J Am Diet Assoc 1988; 88: 1268 1271. 17 Marmot MG, Adelstein AM, Robinson N, Rose G. Changing social class distribution of heart disease. Br Med J 1978; 2: 1109 1112. 18 Klesges RC, Shelton ML, Klesges LM. Effects of television on metabolic rate: potential implications for childhood obesity. Pediatrics 1993; 91: 281 286. 19 Taras HL, Sallis JF, Patterson TL, Nader PR, Nelson JA. Television s influence on children s diet and physical activity. Dev Behav Pediatr 1989; 10; 176 180. 20 Borzekowski DLG, Robinson TN. Effect of commercials on preschoolers. Nutr Res Newslett 2001; 20: 1 2. 21 Robinson TN. Can a school-based intervention to reduce television use decrease adiposity in children grades 3 and 4? Evidence-based Nurs 2000; 3: 43. 22 Epstein LH, Paluch RA, Gordy CC, Dorn J. Decreasing sedentary behaviors in treating pediatric obesity. Arch Pediatr Adolesc Med 2000; 154: 220 226. 23 Robinson TN. Reducing children s television viewing to prevent obesity. JAMA 1999; 282: 1561 1567. 24 Luepker RV, Perry CL, McKinlay SM, Nader PR, Parcel GS, Stone EJ, Webber LS, Elder JP, Feldman HA, Johnson CC. Outcomes of a field trial to improve children s dietary patterns and physical activity. The Child and Adolescent Trial for Cardiovascular Health (CATCH). JAMA 1996; 275: 768 776. 833