BioMed Central

International Journal of Behavioral Nutrition and Physical Activity

ss

Open AcceResearchTelevision viewing and its association with overweight in Colombian children: results from the 2005 National Nutrition Survey: A cross sectional studyLuis F Gomez*1, Diana C Parra1, Felipe Lobelo2, Belen Samper3, José Moreno1, Enrique Jacoby4, Diego I Lucumi1,5, Sandra Matsudo6 and Catalina Borda7

Address: 1Health Division, Fundación FES Social, Bogota, Colombia, 2Department of Exercise Science, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA, 3Department of Evaluation, PROFAMILIA, Bogota, Colombia, 4Non Communicable Diseases Unit, Pan American Health Organization, Washington, D.C., USA, 5Escuela de Medicina. Universidad Pedagógica y Tecnológica de Colombia, Tunja, Colombia, 6Department of Evaluation, Centro de Estudos do Laboratório de Aptidão Física de São Caetano do Sul, São Caetano do Sul, Brazil and 7Research Division, Instituto Colombiano de Bienestar Familiar, Bogota, Colombia

Email: Luis F Gomez* - lfgomez@fundacionfes.org; Diana C Parra - dianacpp79@yahoo.com; Felipe Lobelo - lobelo@mailbox.sc.edu; Belen Samper - belensamper@etb.net.co; José Moreno - josemorenomontoya@gmail.com; Enrique Jacoby - jacobyen@paho.org; Diego I Lucumi - dilucumi@fundacionfes.org; Sandra Matsudo - sandra@celafiscs.org.br; Catalina Borda - Catalina.Borda@icbf.gov.co

* Corresponding author

AbstractBackground: There has been an ongoing discussion about the relationship between time spent watching television andchildhood obesity. This debate has special relevance in the Latin American region were the globalization process hasincreased the availability of screen-based entertainment at home. The aim of this study is to examine the associationbetween television viewing and weight status in Colombian children.

Methods: This cross sectional investigation included children aged 5 to12 yrs from the National Nutrition Survey inColombia (ENSIN 2005). Weight and height were measured in 11,137 children in order to calculate body mass index.Overweight was defined by international standards. Time spent viewing television was determined for these childrenthrough parental reports. Multiple logistic regression analyses were conducted for different subgroups and adjusted forpotential confounders in order to study the association between television viewing and weight status in this population.

Results: Among the surveyed children, 41.5% viewed television less than two hours/day; 36.8% between two and 3.9hours/day and 21.7% four or more hours/day. The prevalence of overweight (obesity inclusive) in this population was11.1%. Children who were classified as excessive television viewers (between two and 3.9 hours/day or 4 or more hours/day) were more likely to be overweight (OR: 1.44 95% CI: 1.41–1.47 and OR: 1.32 95% CI: 1.30–1.34, respectively) thanchildren who reported to watch television less than 2 hours/day. Stratified analyses by age, gender and urbanization levelsshowed similar results.

Conclusion: Television viewing was positively associated with the presence of overweight in Colombian children. Apositive association between urbanization level and television viewing was detected. Considering that the majority ofColombian children lives in densely populated cities and appear to engage in excessive television viewing these findingsare of public health relevance for the prevention of childhood obesity.

Published: 19 September 2007

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 doi:10.1186/1479-5868-4-41

Received: 29 December 2006Accepted: 19 September 2007

This article is available from: http://www.ijbnpa.org/content/4/1/41

© 2007 Gomez et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 1 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

BackgroundChildhood obesity is associated with the presence of car-diovascular disease risk factors such as hyperlipidemia,hypertension, glucose intolerance, as well as their cluster-ing [1]. Moreover, overweight children and adolescentsare at increased risk for obesity and related cardio-meta-bolic consequences during adulthood, thus entailing thatobesity prevention should begin during childhood [2,3].

The prevalence of childhood obesity has reached epi-demic proportions in developed nations [4]. Recent datafrom the United States indicates that 37.2% of children 6to 11 years old are overweight [5]. In developing countriessuch as those located in the Latin American region, child-hood obesity is also emerging as a public health threat[6,7]. For example, in 1998 the prevalence of overweightamong 5–11 y old Mexican children reached 19.5% [8].Nevertheless, in a region where childhood malnutritioncontinues to be a top public health priority, data regardingthe prevalence and trends of childhood obesity are scarce[9]. In addition, information about correlates of child-hood obesity such as excessive TV viewing or decreasedphysical activity levels is limited.

Rapid processes of industrialization coupled with acceler-ated epidemiologic and nutritional transitions have beenlinked to the rising prevalence of childhood obesity indeveloping countries [10-15]. In this context, increase intime devoted to sedentary activities and reduced levels ofphysical activity, in conjunction with energy-dense diets,are considered factors contributing to its development[16]. In particular, television (TV) viewing has been exten-sively studied in relation to fatness, mainly among chil-dren and adolescents from developed countries [17,18].The available evidence indicates that there is a significantassociation between TV viewing and obesity among chil-dren [18-21].

Despite the fact that increases in the prevalence of child-hood obesity have been documented worldwide and thattime devoted to TV viewing has been implicated in thesetrends, there is little information regarding the associationbetween TV viewing and obesity in Latin America children[22]. This information has special relevance in a develop-ing country like Colombia where the globalization proc-ess has increased the availability of television, computersand video games at home.

Consequently, the aim of this study is to examine theassociation between TV viewing and overweight (obesityinclusive) in a representative sample of 5–12 y oldColombian children living in both rural and urban areas.

MethodsStudy designThe present investigation used data from the NationalNutrition Survey in Colombia (in Spanish: EncuestaNacional de la Situación Nutricional en Colombia, ENSIN2005), which was conducted by the Family WelfareColombian Institute (in Spanish: Instituto Colombianode Bienestar Familiar, ICBF) with the logistic support ofthe Association for the Colombian Family Wellness (inSpanish: Asociación Probienestar de la Familia Colombiana,PROFAMILIA) [23]. A stratified, probabilistic, multi-stage,cluster sampling of households was performed to obtainnational and sub-regional representativeness (16 sub-regions), with oversampling of rural areas and low socio-economic status (SES) groups. Logistic aspects and techni-cal details of ENSIN can be found elsewhere [23]. Themain objective of this national survey was to estimate theprevalence of nutritional problems and selected healthconditions in the Colombian population. For the presentstudy, data from 11,137 children between 5 and 12 yearsold (37% of the ENSIN sample) with complete informa-tion about television viewing, weight status and covariateswas used. Complete information was available for 80.2%of the surveyed children. Using a propensity score weidentified that those children with incomplete informa-tion had significant differences in the distribution of soci-oeconomic status and television viewing, however thesewere of low magnitude and the investigators judged thatthe study's internal validity will not be affected.

Seventy nutritionists received a standardized trainingprior to data collection with 62 of the 70 being selected asinterviewers. Thirteen nutritionists were also specificallytrained to take anthropometric measures.

Measurement of the dependent variableWeight and height were taken directly at the children'shomes using standardized measuring equipment andwith children in light clothing and shoes removed. Bodyweight was measured to the nearest 0.1 kg using a cali-brated digital scale (SECA model 770) which was properlyadjusted for the Colombian geographical latitude. Heightwas measured to the nearest 1.0 cm using a portable sta-diometer (Shorr Productions). BMI was computed bydividing body weight by height squared (kg/m2). Over-weight (obesity inclusive) was defined using the genderand age-specific cut-off points for BMI adopted by theInternational Obesity Task Force [24]. These cut-offs werederived by averaging the percentiles of six countries(including Brazil representing the Latin-American popula-tion), equivalent to a BMI of 25 kg/m2 at age 18. This clas-sification was selected for the present study because itconstitutes the international standard to define childhoodoverweight, enabling cross-country comparisons.

Page 2 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

Measurement of TV viewingTV viewing and video game use were determined by ask-ing the following question: "During the last seven days, did__________ (child's name) watch television or play videogames?" Informers who provided a positive answer werealso asked about the frequency of the event: "How manydays?" followed by "How much time did _________ (child'sname) usually spend during one of those days watching televi-sion or playing video games?". If the interviewed parent orguardian could not answer the last question due to thevariability of the report from day to day, the interviewerwould then ask: "What is the total amount of time that__________ (child's name) spent over the last seven dayswatching television or playing video games?".

Based on the combined TV viewing and video-game useinformation, children were then classified into threegroups: less than two hours per day, 2 to 3.9 hours perday, and four or more hours per day. Current publichealth and clinical recommendations regarding TV view-ing in the pediatric setting informed the selection of thisthreshold (≥ 2 hours/day) to study its effects on children'sweight status [25,26].

For the purposes of this study, we only refer to TV viewingsince the majority of the children belonged to low SESgroups and did not have access to video-games.

Measurement of covariatesThe following covariates were included, age (stratified intwo groups 5 to 8 years and 9 to 12 years), gender, urban-ization levels, and socioeconomic strata (SES) of the fam-ily. Based on their population density, cities and townsincluded in the study were divided into four urbanizationlevels. Urbanization level I included mostly rural settle-ments with 10,000 inhabitants or less. Urbanization levelII included cities or towns with 10,001 to 30,000 inhabit-ants. Urbanization level III were cities with 30,001 to100,000 inhabitants and urbanization level IV were citieswith more than 100,000 inhabitants.

The SES of the child's family was determined according tothe national SISBEN index which takes into accountsocio-demographic characteristics (family composition,employment status, family income, and education level),living conditions (construction type and materials), andaccess to public utilities (sewer, electricity, potable waterand garbage collection) [27]. Based on this information,six levels are defined with one being the poorest and sixbeing the wealthiest. For this study, levels 3 to 6 were col-lapsed into one group to improve the efficiency of theanalyses.

Statistical analysisThe distribution of selected covariates by TV viewinggroup was evaluated using Pearsons' Chi square test. Theassociation between TV viewing and overweight (obesityinclusive) was assessed using multiple logistic regressionanalysis while adjusting for potential confounders (gen-der, age groups, SES and urbanization levels). In addition,stratified analyses were performed by age, gender, urbani-zation level, and SES of the family. Collinearity was exam-ined using regression diagnostic tests including varianceinflation factor (VIF) with their values always less than0.99 [28]. The analysis took into account an unequalselection of probabilities resulting from the complex sam-pling design. All the statistical analyses were performedusing STATA while considering sampling strata, sampleunits and weights [29].

ResultsStudy population characteristicsGender distribution for the study sample (n = 11,137) wasapproximately equal (49.9% males) and the mean agewas 8.5 y (SD = 2.28) with 50.3% of the children being inthe 9–12 y old range. Approximately 41.1% of the chil-dren were in the lowest SES. Most of the surveyed children(71.3%). lived in rural areas and urban settlements withpopulations of less than 10.001 inhabitants

Prevalence of TV viewingFor the study population 41.5% of parents reported theirchild's TV viewing level to be < 2 hours/day, 36.8%between 2 and 3.9 hours/day, and 21.7% to be 4 or morehours/day. Excessive TV viewing (≥ 2 hours/day) wasmore prevalent among older children (62.6%; p < 0.001),males (60.6%; p < 0.001), those with middle to high SES(levels 3 – 6; 66.5%; p < 0.001), and in children living inthe most urbanized areas (level IV; 71.3 %; p < 0.001)(Table 1).

Prevalence of overweightOf the studied children, 11.1% were classified as over-weight (obesity inclusive) with a higher prevalenceamong younger children (5–8 y: 11.3%; p < 0.001), thoseliving in the most urbanized areas (level III and IV: 14.1%and 14.6% respectively; p < 0.001), and those in middleto high SES level (levels 3–6: 17.2 %; p < 0.001) (Table 2).

Associations between TV viewing and overweightIn the logistic regression models, children classified asexcessive TV viewers (2 to 3.9 hours/day or 4 or morehours/day) were more likely to be overweight (OR: 1.4495% CI: 1.41–1.47 and OR: 1.32 95% CI: 1.30–1.34,respectively) than those who watched less than 2 hours/day. This pattern of association was also found in analysisstratified by gender, age and urbanization levels. (Table 3)

Page 3 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

In addition, the strength of association between TV view-ing and weight status was greater for those in the 2 – 3.9hours/day group than for those in the 4 or more hours/day group, with the exception of the 5 to 8 y old group andthose who lived in rural areas and urban settlements ofless than 10,001 inhabitants.

DiscussionIn this study we found a positive association betweenexcessive TV viewing and the presence of overweightamong Colombian children. This association wasobserved for children of different gender, age groups (5–8and 9–12 y old) and for children living in different urban-ization levels.

The magnitude of the associations detected between TVviewing groups and weight status ranged from (OR) 1.32to 1.44 for the total sample and from 1.19 to 1.78 in thesub-group analyses. Previous studies exploring the associ-ation between electronic media use and health-relatedoutcomes have found conflicting results. Some investiga-tions have not detected significant associations [30,31]but most have reported a positive relationship [17,32-36].The association between TV viewing and childhood obes-ity has been found consistently across different sociode-mographic groups, as explored by studies conductedamong different age groups, [32,33,35,37] and in chil-dren from developed[17] and developing countries [22].

Table 2: Prevalence of overweight (obesity inclusive) by selected sociodemographic characteristics and television viewing among 11,137 children aged 5 to 12 years. Analysis conducted from the National Nutrition Survey (ENSIN) Colombia. 2005

Characteristics P* SE p

All participants 11.1 0.05Age groups, yrs5–8 11.3 0.08 < 0.0019–12 10.9 0.06SexMale 10.0 0.08 < 0.001Female 12.2 0.04Socio Economic StatusLevel 1 (Lowest) 5.0 0.08 < 0.001Level 2 11.8 0.06Levels 3 to 6 (Middle-high) 17.2 0.05Levels of urbanizationLevel I (rural an urban areas with 10.000 inhs or less)

8.5 0.07 < 0.001

Level II (10.001 to 30.000 inhs) 11.6 0.03Level III (30.001 to 100.000 inhs) 14.1 0.09Level IV (more than 100.000 inhs) 14.6 0.01Television viewing exposure levels< 2 hrs/day 8.5 0.05 < 0.0012 – 3.9 hrs/day 13.5 0.104 or more hrs/day 13.0 0.05

*Weighted percentage (P) and standard error (SE)

Table 1: Prevalence of television viewing by sociodemographic characteristics among 11,137 children aged 5 to 12 years. Analysis conducted from the National Nutrition Survey (ENSIN) Colombia. 2005

Characteristics n* < 2 hrs/day 2 – 3.9 hrs/day ≥ 4 hrs/day p

P** SE P** SE P** SE

All participants 11,137 41.5 0.03 36.8 0.02 21.7 0.02Age groups, yrs5–8 5,539 45.7 0.04 34.8 0.02 19.5 0.02 < 0.0019–12 5,598 37.4 0.03 38.7 0.03 23.9 0.02SexMale 5,568 39.4 0.05 38.1 0.03 22.5 0.03 < 0.001Female 5,569 43.6 0.03 35.4 0.02 21.0 0.02Socio Economic StatusLevel 1 (Lowest) 4,576 56.0 0.04 28.9 0.03 15.1 0.02 < 0.001Level 2 4,124 34.7 0.04 41.1 0.03 24.2 0.03Level 3 to 6 (Middle – high) 2,437 33.5 0.05 40.3 0.04 26.2 0.03Levels of urbanizationLevel I (rural and urban areas with 10.00 inhs or less)

7,946 51.1 0.04 32.9 0.03 16.0 0.03 < 0.001

Level II (10.001 to 30.000 inhs) 833 30.4 0.01 42.1 0.01 27.5 0.01Level III (30.001 to 100.000 inhs) 1,292 33.7 0.08 39.4 0.07 26.9 0.03Level IV (more than 100.000 inhs) 1,066 28.7 0.01 42.1 0.01 29.2 0.01

*Unweighted sample size.**Weighted percentage (P) and standard error (SE)

Page 4 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

Furthermore, a recent meta-analysis found a positive, yetweak relationship, between fatness and TV viewing [18].The findings of the present study are in concordance withprevious evidence on the topic and have important publichealth implications [38].

Among children aged 5–8 yrs and children living in ruralor urban areas with less than 10,001 inhabitants, wefound increasing odds for overweight as TV viewingincreased. The stability of the estimates we obtained forthe association between weight status and TV viewing mayhave been affected by information biases related to thislast variable. For example, a recent study indicates thathaving a TV in the children's bedroom should become animportant control variable in studies where estimates areobtained by proxy report [39]. This is because parentsoverestimate children's TV viewing time when no TV ispresent in the bedroom but underestimate it when thereis a TV set in the bedroom [39]. In this study we did not

have this information to adjust for, and determine thedirection and magnitude of the error in the exposure var-iable.

Parental reports of TV viewing time in older children (9–12 y old) might not be as adequate because these childrenare more independent than their younger counterpartsand their parents could have less control over their dailyactivities, thus overestimating their TV viewing time.

Similarly, the quality of information regarding TV viewingtime might be better for children of low SES living in ruralor small urban settlements where, in most of the cases,there is only one TV. In such cases parents will probablyhave a better idea of their children's TV viewing time.More detailed information on the children's TV viewingpatterns will be an important addition to future studiesexploring health-related effects of TV viewing and otherelectronic entertainment. While the effect of excessive TV

Table 3: Odds Ratios for overweight (obesity inclusive) by sex, age groups, and urbanization levels among 11,137 children aged 5 to 12 years, Analysis conducted from the National Nutrition Survey (ENSIN) Colombia. 2005

Television viewing exposure levels

Adjusted OR 95% CI p

a Total population< 2 hrs/day 1.00 Referent2 – 3.9 hrs/day 1.44 (1.41–1.47) < 0.001≥ 4 hrs/day 1.32 (1.30–1.34) < 0.001

b All males< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.51 (1.46–1.56) < 0.001≥4 hrs/day 1.38 (1.34–1.42) < 0.001

b All females< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.39 (1.34–1.44) < 0.001≥ 4 hrs/day 1.29 (1.26–1.32) < 0.001

c All children aged 5 to 8 ys< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.19 (1.15–1.22) < 0.001≥ 4 hrs/day 1.32 (1.29–1.36) < 0.001

c All children aged 9 to 12 ys< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.78 (1.73–1.84) < 0.001≥ 4 hrs/day 1.40 (1.36–1.43) < 0.001

d Level I (rural and urban areas with 10000 inhabitants or less)< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.29 (1.24–1.34) < 0.001≥ 4 hrs/day 1.49 (1.46–1.52) < 0.001

d Level II to IV (10001 inhabitants and more)< 2 hrs/day 1.00 Referent2 -3.9 hrs/day 1.56 (1.53–1.60) < 0.001≥ 4 hrs/day 1.26 (1.24–1.29) < 0.001

a Adjusted by sex, age groups, SES and urbanization levels.b Adjusted by age groups, SES and urbanization levels.c Adjusted by sex, SES and urbanization levels.d Adjusted by sex, age groups and SES.

Page 5 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

viewing time on adiposity could be explained by the dis-placement of physically active pursuits and/or increases inenergy intake while viewing TV, [40] an important ques-tion is how these factors behave differently depending onthe cultural and social contexts. Social and physical envi-ronments have been recognized as significant determi-nants of health behaviors [40,41]. Aspects such as livingin a safe place that supports physical activity play animportant role in the prevention of childhood obesity[41]. It may be the case that families living in low SESareas with high crime rates, choose to restrict children'soutdoor play and replace this time with indoor activitiessuch as TV viewing [42]. Likewise, food advertising andthe availability of fast food products are important factorsto consider in the development of childhood obesity inthe Latin American region [43]. The availability of energy-dense foods in developing countries has certainlyincreased as a result of the industrialization and globaliza-tion processes [44]. Further studies will need to explorethe extent to which these products are being advertised tochildren and youth and whether an association existsbetween food advertisement and childhood obesity inLatin America. All of these aspects should be contem-plated in future studies in order to better understand themyriad of interacting factors driving the relationshipbetween TV viewing and childhood obesity in particularfor Latin American youth.

The findings of this study point to the existence of a posi-tive association between the level of urbanization withboth TV viewing time and prevalence of childhood obes-ity. The fact that the most urbanized areas have moreaccess to television as an indirect result of the globaliza-tion process might explain these results. The Latin Ameri-can region is urbanizing at an extremely rapid pace.Currently, 75% of the population lives in urbanized areasand it is predicted that this proportion will increase to84% by 2030 [45]. As most Latin American countriesexperience rapid epidemiological and nutritional transi-tions it is important to recognize that increases in theprevalence of childhood obesity occur as a function ofthese epidemiological and societal trends of development[46]. Therefore, children living in urban areas andexposed to excessive TV viewing should be considered agroup at particular risk for childhood obesity.

Several strengths can be identified in this study. To ourknowledge, this is the first investigation to explore theassociation between TV viewing and overweight in anational representative sample of children from a LatinAmerican country. The use of data from a national surveyallowed us to contrast the association between TV viewingand weight status by children living in different urbaniza-tion levels and SES. The study previously conducted inMexico was restricted to children who resided in the Fed-

eral District [22]. On the other hand, some limitationsneed to be discussed. The methodology used for the quan-tification of TV viewing might be affected by informationbiases related to inaccuracies in the parental reports of TVwatching. Distortions such as lack of awareness of theirchild's TV viewing habits, recall bias, desire to please, andreluctance to tell the truth might be present. Anotheraspect to be considered is the difficulty to administer sur-veys in low educated populations from developing coun-tries. Nevertheless, proxy reports are the most usedmethod (and possibly the only feasible one) to measureTV viewing time in a population survey of childrenbetween 5 and 12 y old [47]. These issues emphasize theneed of improving the design of instruments to bettermeasure TV viewing time in ways that meet the specificneeds of Latin America populations. In addition, thecross-sectional nature of this study does not allow toestablish a causal relationship between TV viewing andchildren's weight status. Despite these limitations, theresults of this study should promote interventions to limitTV viewing as part of childhood obesity prevention pro-grams in the region.

ConclusionTime spent in sedentary activities in particular TV viewingis emerging as a risk factor for different health outcomesin children including obesity and related cardio-meta-bolic disarrays. There is a growing concern that rapid proc-esses of globalization and urbanization occurring indeveloping regions like Latin America will impact theprevalence of childhood obesity in part as a result ofincreased availability of television at home. Our findingssuggest that TV viewing is positively associated with thepresence of overweight and obesity among Colombianchildren. At the same time, both TV viewing and the prev-alence of overweight were found to be positively associ-ated with urbanization level. The public healthimplications of these findings are considerable, since themajority of Colombian children live in densely populatedareas and engage in excessive TV viewing. These trends area call for action for the development and implementationof programs and national policies for the prevention ofchildhood obesity in Colombia and the Latin Americanregion. An important component of such programsshould be strategies to limit TV viewing among childrenand adolescents

Competing interestsThe author(s) declare that they have no competing inter-ests.

Authors' contributionsLFG conceived the study and participated in the analysisand drafting of the manuscript. DP and FL participated inthe analyses and interpretation of data and in the drafting

Page 6 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

of the manuscript. BS, JM and DL provided valuable com-ments in the analyses and discussion sections of the man-uscript. EJ, SM, CB contributed in data collection protocolfor ENSIN and provided valuable comments in the draft-ing of the manuscript. All authors read and approved thefinal manuscript

AcknowledgementsThis work was supported by the Family Welfare Colombian Institute and PROFAMILIA. The authors are grateful for the valuable comments pro-vided by Gabriel Ojeda, Luz Mariela Manjarres and Orlando Scoppetta.

References1. Goodman E, Dolan LM, Morrison JA, Daniels SR: Factor analysis of

clustered cardiovascular risks in adolescence: obesity is thepredominant correlate of risk among youth. Circulation 2005,111:1970-1977.

2. Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH: Predict-ing obesity in young adulthood from childhood and parentalobesity. N Engl J Med 1997, 337:869-873.

3. Janssen I, Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM, Bou-chard C, Berenson GS: Utility of childhood BMI in the predic-tion of adulthood disease: comparison of national andinternational references. Obes Res 2005, 13:1106-1115.

4. Malecka-Tendera E, Mazur A: Childhood obesity: a pandemic ofthe twenty-first century. Int J Obes (Lond) 2006, 30(Suppl2):S1-S3.

5. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, FlegalKM: Prevalence of overweight and obesity in the UnitedStates, 1999–2004. JAMA 2006, 295:1549-1555.

6. De Onis M, Blossner M: Prevalence and trends of overweightamong preschool children in developing countries. Am J ClinNutr 2000, 72:1032-1039.

7. Wang Y, Monteiro C, Popkin BM: Trends of obesity and under-weight in older children and adolescents in the UnitedStates, Brazil, China, and Russia. Am J Clin Nutr 2002,75:971-977.

8. Hernandez B, Cuevas-Nasu L, Shamah-Levy T, Monterrubio EA, Ram-írez-Silva CI, Garcia-Feregrino R, Rivera JA, Sepúlveda-Amor J: Fac-tores asociados con sobrepeso y obesidad en niñosmexicanos de edad escolar: Resultados de la EncuestaNacional de Nutrición 1999. Salud Publica Mex 2003,45:s551-s557.

9. Prentice AM: The emerging epidemic of obesity in developingcountries. Int J Epidemiol 2006, 35:93-99.

10. Rivera JA, Barquera S, Gonzalez-Cossio T, Olaiz G, Sepulveda J:Nutrition transition in Mexico and in other Latin Americancountries. Nutr Rev 2004, 62:S149-S157.

11. Popkin BM: The nutrition transition and obesity in the devel-oping world. J Nutr 2001, 131:871S-873S.

12. Jenovesi JF, Bracco MM, Colougnati FA, Taddei JA: Physical activityprofile among children of public school of different nutri-tional status. Brazilian Journal of Movement Science 2003, 11:57-62.

13. Uauy R, Albala C, Kain J: Obesity trends in Latin America: tran-siting from under- to overweight. J Nutr 2001, 131:893S-899S.

14. Albala C, Vio F, Kain J, Uauy R: Nutrition transition in Chile:determinants and consequences. Public Health Nutr 2002,5:123-8.

15. Rissin A: [Nutritional transition in Brazil: geographic andtemporal trends]. Cad Saude Publica 2003, 19(Suppl 1):S181-91.

16. Daniels SR, Arnett DK, Eckel RH, Gidding SS, Hayman LL, KumanyikaS, Robinson TN, Scott BJ, St Jeor S, Willians CL: Overweight in chil-dren and adolescents: pathophysiology, consequences, pre-vention, and treatment. Circulation 2005, 111:1999-2012.

17. Andersen RE, Crespo CJ, Bartlett SJ, Cheskin LJ, Pratt M: Relation-ship of Physical Activity and Television Watching With BodyWeight and Level of Fatness Among Children: Results Fromthe Third National Health and Nutrition Examination Sur-vey. JAMA 1998, 279:938-942.

18. Marshall SJ, Biddle SJ, Gorely T, Cameron N, Murdey I: Relation-ships between media use, body fatness and physical activity

in children and youth: a meta-analysis. Int J Obes Relat MetabDisord 2004, 28:1238-1246.

19. Hancox RJ, Milne BJ, Poulton R: Association between child andadolescent television viewing and adult health: a longitudinalbirth cohort study. Lancet 2004, 364:257-262.

20. Epstein LH, Paluch RA, Gordy CC, Dorn J: Decreasing SedentaryBehaviors in Treating Pediatric Obesity. Arch Pediatr AdolescMed 2000, 154:220-226.

21. Robinson TN: Reducing Children's Television Viewing to Pre-vent Obesity: A Randomized Controlled Trial. JAMA 1999,282:1561-1567.

22. 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 amongchildren in Mexico city. Int J Obes Relat Metab Disord 1999,23:845-854.

23. Instituto Colombiano de Bienestar Familiar: Encuesta Nacional dela Situación Nutricional en Colombia. 2005.

24. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH: Establishing a standarddefinition for child overweight and obesity worldwide: inter-national survey. BMJ 2000, 320:1240-1243.

25. US Department of Health and Human Services: Healthy People 20102nd edition. Washington, D.C.: US Government Printing Office; 2000.

26. American Academy of Pediatrics CoPE: Children, Adolescentsand Television. Pediatrics 2001, 107:423-426.

27. DANE: Departamento Administrativo Nacional de Estadísti-cas. Documentos Técnicos Sobre Mercado Laboral. 2003[http://www.dane.gov.co/files/investigaciones/empleo/]. Acceded Sep-tember 19 2006

28. Kleinbaum D, Kupper L, Muller K: Applied regression analysis and othermultivariate methods 2nd edition. Boston: PWS-Kent; 1987.

29. Stata Corporation: College Station: Statistical Software: release 9. Texas2005.

30. Robinson TN, Hammer LD, Killen JD, Kraemer HC, Wilson DM,Hayward C, Taylor CB: Does television viewing increase obes-ity and reduce physical activity? Cross-sectional and longitu-dinal analyses among adolescent girls. Pediatrics 1993,91:273-280.

31. Wolf AM, Gortmaker SL, Cheung L, Gray HM, Herzog DB, ColditzGA: Activity, inactivity, and obesity: racial, ethnic, and agedifferences among schoolgirls. Am J Public Health 1993,83:1625-1627.

32. Davison KK, Marshall SJ, Birch LL: Cross-sectional and longitudi-nal associations between TV viewing and girls' body massindex, overweight status, and percentage of body fat. J Pediatr2006, 149:32-37.

33. Kaur H, Choi WS, Mayo MS, Harris KJ: Duration of televisionwatching is associated with increased body mass index. J Pedi-atr 2003, 143:506-511.

34. Proctor MH, Moore LL, Gao D, Cupples LA, Bradlee ML, Hood MY,Ellison RC: Television viewing and change in body fat frompreschool to early adolescence: The Framingham Children'sStudy. Int J Obes Relat Metab Disord 2003, 27:827-833.

35. Armstrong CA, Sallis JF, Alcaraz JE, Kolody B, McKenzie TL, HovellMF: Children's television viewing, body fat, and physical fit-ness. Am J Health Promot 1998, 12:363-368.

36. Ekelund U, Brage S, Froberg K, Harro M, Anderssen SA, Sardinha LB,Riddoch C, Andersen LB: TV Viewing and Physical Activity AreIndependently Associated with Metabolic Risk in Children:The European Youth Heart Study. PLoS Med 2006, 3:e488.

37. DuRant RH, Baranowski T, Johnson M, Thompson WO: The rela-tionship among television watching, physical activity, andbody composition of young children. Pediatrics 1994,94:449-455.

38. Hancox RJ, Poulton R: Watching television is associated withchildhood obesity: but is it clinically important? Int J Obes(Lond) 2006, 30:171-175.

39. Robinson JL, Winiewicz DD, Fuerch JH, Roemmich JN, Epstein LH:Relationship between parental estimate and an objectivemeasure of child television watching. Int J Behav Nutr Phys Act2006, 3:43.

40. Robinson TN: Television viewing and childhood obesity. PediatrClin North Am 2001, 48:1017-1025.

41. Nelson MC, Gordon-Larsen P, Song Y, Popkin BM: Built and socialenvironments associations with adolescent overweight andactivity. Am J Prev Med 2006, 31:109-117.

Page 7 of 8(page number not for citation purposes)

International Journal of Behavioral Nutrition and Physical Activity 2007, 4:41 http://www.ijbnpa.org/content/4/1/41

Publish with BioMed Central and every scientist can read your work free of charge

"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."

Sir Paul Nurse, Cancer Research UK

Your research papers will be:

available free of charge to the entire biomedical community

peer reviewed and published immediately upon acceptance

cited in PubMed and archived on PubMed Central

yours — you keep the copyright

Submit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.asp

BioMedcentral

42. Fraser B: Latin America's urbanisation is boosting obesity.Lancet 2005, 365(9476):1995-1996.

43. Hawkes C: Uneven dietary development: linking the policiesand processes of globalization with the nutrition transition,obesity and diet-related chronic diseases. Global Health 2006,2:4.

44. Adair LS, Popkin BM: Are child eating patterns being trans-formed globally? Obes Res 2005, 13:1281-1299.

45. Cohen B: Urban growth in developing countries: a review ofcurrent trends and a caution regarding existing forecasts.World Dev 2004, 32:23-51.

46. Popkin BM, Gordon-Larsen P: The nutrition transition: world-wide obesity dynamics and their determinants. Int J Obes RelatMetab Disord 2004, 28(Suppl 3):S2-S9.

47. Marshall SJ, Gorely T, Biddle SJ: A descriptive epidemiology ofscreen-based media use in youth: a review and critique. J Ado-lesc 2006, 29:333-349.

Page 8 of 8(page number not for citation purposes)