%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
There is little doubt that lack of physical activity in children constitutes a risk factor for childhood overweight/obesity(Reference Moore, Nguyen, Rothman, Cupples and Ellison1, Reference Must and Tybor2). Indeed, there are plenty of studies reporting a positive correlation between hours of watching television (TV) and childhood overweight and obesity(Reference Mendoza, Zimmerman and Christakis3–Reference Bryant, Lucove, Evenson and Marshall5). The number of hours spent in front of the TV set, however, reflects physical inactivity rather than physical activity. Physical activity, an important determinant of energy balance and therefore likely to influence obesity risk, is promoted in many prevention programmes but was only rarely measured on an individual level in the evaluation of these programmes. Several studies attempting to measure physical activity in children failed to show a significant association with overweight/obesity(Reference Jiang, Rosenqvist, Wang, Greiner, Ma and Toschke6–Reference Aeberli, Kaspar and Zimmermann8). This situation may be due to difficulties in measurement (especially in pre-school children)(Reference Campbell and Hesketh9) that arise even when technical equipment (accelerometry) is used(Reference Toschke, von Kries, Rosenfeld and Toschke10). Only a few studies carried out in this age group make use of methods other than accelerometry(Reference McGarvey, Keller, Forrester, Williams, Seward and Suttle11–Reference Burdette, Whitaker and Daniels13).
In the present study, we report on a simple approach to assess physical activity in pre-school children based on a set of questions in parental questionnaires to classify children as physically active or not. The primary objective of the study was to test whether this classification was associated with childhood overweight/obesity. The secondary objective was to assess whether this classification was also correlated with the results of a simple motor test.
Methods
Study population
The study was conducted within the framework of the first and second survey of the health monitoring units in Bavaria (GME, Gesundheits-Monitoring-Einheiten)(Reference Bolte, Heissenhuber, von Kries, Liebl, Zapf, Wildner and Fromme14). GME, established in 2004, is supported by the Bavarian State Ministry of the Environment, Public Health and Consumer Protection and is managed by the Bavarian Health and Food Safety Authority. Its aims are: (i) the epidemiological surveillance of health and environmental factors with special emphasis on children; (ii) the prompt provision of results to political leaders and decision makers; and (iii) the testing and evaluation of intervention and prevention strategies in cooperation with external research partners. To facilitate the last aim a modular approach is employed. The six regions covered are selected in order to get data representative for Bavaria. The study was approved by the ethics committee of the Bavarian medical council (BLÄK, Bayerische Landesärztekammer). All participating parents gave their written consent. In three urban and three rural Bavarian regions, 8127 and 8558 children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, respectively, were invited to participate.
Measurements
Height and weight were measured using standard stadiometers and calibrated scales (usually digital) by trained staff of the local health authority, a motor test was conducted and questionnaires were handed out to parents. Variables relevant to the present study are defined as follows.
1. Overweight/obesity: age- and sex-specific BMI cut-off values were used to classify children as overweight or obese according to Cole’s reference system(Reference Cole, Bellizzi, Flegal and Dietz15).
2. Side-to-side jumps (SHH): the number of jumps over a bar a child can perform within 15 s. The child should jump and land with both feet simultaneously. The numbers of two runs were added up. This test is one item of the Karlsruher-Motorik-Screening(Reference Bös, Bappert, Tittlbach and Woll16), a four-item screener designed for pre-school children to assess physical fitness in terms of strength, coordination and motility. After a pilot phase of 3 years, norm values were calculated on the basis of a sample consisting of 1288 children.
Physical activity (from parent questionnaire)
Children’s activities such as ‘climbing’, ‘playing with a ball’, ‘playing tag’, ‘cycling using a bicycle or kickboard’ and ‘skating’ were listed. Parents were asked to rate the frequency of their child’s taking part in these activities as ‘never’, ‘less than once per week’, ‘1–2 times per week’, ‘3–6 times per week’, ‘every day’ or ‘don’t know’. The following cut-off points were chosen based on the frequency distribution of answers in order to divide them into two groups of roughly equal size: (i) climbing, playing with a ball and playing tag were rated as 0 (=rarely) if reported as less than 1–2 times per week, otherwise as 1 (=frequently); (ii) cycling was rated as 0 (=rarely) if reported as less than 3–6 times per week, otherwise as 1 (=frequently); and (iii) skating was rated as 0 (=rarely) if reported as never, otherwise as 1 (=frequently). Out of these dichotomized answers the variable physical activity was compiled and set to 1 (=yes) if at least three of the above activities were rated as frequent. Conversely, if at least three of the five answers were rated as not frequent, the variable was set to 0 (=no).
Further questions on how the child usually goes to kindergarten, to friends, to the playground, etc. (on foot, by bicycle, by car, etc.), whether there is free space to play outside nearby and TV/computer consumption were asked and dichotomized. Classification as ‘rare TV watching’ required answers of <1 and 3 h of TV/computer use per day during the week and weekend, respectively, and no TV set in the child’s bedroom.
Parental education (rated as high if at least one of the parents had passed the secondary school qualification examination after 10 years (‘Realschulabschluss’) or had a higher level of education), parental BMI, German nationality, premature birth and complications after birth were recorded as potential confounders. Parental BMI was calculated from reported weight and height of the father; if not available, of the mother. We chose this approach because paternal BMI has a slightly stronger effect on children’s weight status and including both parents’ BMI would have led to considerably more missing values.
Statistical analysis
Statistics were computed in the SAS statistical software package version 9·1 (SAS Institute, Cary, NC, USA). For proportions, exact (binomial) confidence intervals are given. Multivariate logistic regression analysis in conjunction with a backward elimination algorithm (at α = 0·05) was used to model odds ratios for overweight/obesity; Poisson regression was applied to model the multiple of the number of side-to-side jumps in the exposed compared with the non-exposed group.
Results
The proportion of completed questionnaires was 78 % and 73 % in the first and second survey, respectively, yielding 12 556 informative participants available for further analysis. For 12 199 (97·2 %) of these, weight, height and gender were reported allowing for classification of overweight or obesity. Measurements of the outcome SHH were available for 10 696 children (94·7 %, smaller denominator, since SHH was omitted in the Munich region for the second survey). Population characteristics are presented in Table 1. Boys were rated as physically active significantly more frequently (41·2 %) than girls (35·6 %).
Characteristics of the study population: subset of children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, Bavaria, Germany
The prevalences of overweight and obesity were lower in children for whom high physical activity was reported. Table 2 shows the prevalences of overweight and obesity by other variables which possibly influence BMI. All variables that were significant in bivariate analysis entered the first multivariate logistic regression model. The final model revealed adjusted effects for physical activity on overweight and obesity (Table 3) with almost identical OR compared with the bivariate analysis (bivariate OR of 0·803 and 0·672 for overweight and obesity, respectively).
Prevalences of overweight and obesity by reported physical activity and other variables: subset of children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, Bavaria, Germany
Significant differences are typeset in bold face.
Effect estimatesFootnote * on overweight and obesity from multivariate logistic regression analysis with P values from type III analysis (n 8712): subset of children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, Bavaria, Germany
* Adjusted for all other explanatory variables in the table.
As shown in Table 4, children reported as being physically active performed significantly more side-to-side jumps than those reported as not being physically active. Other explanatory variables showing significant associations with SHH in bivariate analysis were ‘sex’, ‘free space to play outside’, ‘physically active’, ‘high parental education’, ‘German nationality’, ‘premature birth’ and ‘complications within 4 weeks after birth’. The highest impact on SHH was found for reported physical activity. All significant variables from the bivariate analysis were included in the Poisson regression model (Table 5). The obtained effect estimate for physical activity fitted very well with the difference seen in bivariate analysis. Since there were also considerable differences in SHH between regions (data not shown), these were further added to the model. The variable ‘free space to play outside’, which was only marginally significant in bivariate analysis (P = 0·0592 using the non-parametric Wilcoxon–Mann–Whitney test), was removed (P = 0·2736 in multivariate analysis), resulting in virtually no change of the other effect estimates.
Results of motor testing by different explanatory variables: subset of children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, Bavaria, Germany
Significant differences are typeset in bold face.
AdjustedFootnote * effects of predictors that were significant in the bivariate analysis on the outcome variable SHH (side-to-side jumps) (n 8403): subset of children taking part in the obligatory school entrance health examination in 2004–5 and 2005–6, Bavaria, Germany
* Adjusted for all other explanatory variables in the table.
† The factors reflect direction (greater or less than 1) and strength of the effect. The number of jumps for children reported e.g. as physically active has to be multiplied by the factor 1·070, implying that these children achieve 7·0 % more jumps on average.
Discussion
Data collected in pre-school children in three urban and three rural Bavarian regions revealed that children with high physical activity assessed in a parental questionnaire showed a lower prevalence of overweight and obesity and better results in motor testing.
Energy expenditure and therefore physical activity plays a crucial role for the development of overweight. This might be particularly relevant for pre-school children, who show a higher level of physical activity compared with other age groups(Reference Sigmund, De Ste Croix, Miklánková and Frömel17). Surprisingly, there are few studies on the interdependencies of physical activity in pre-school children and obesity. Additionally, these studies have yielded conflicting results with only a few reporting significant associations(Reference Jago, Baranowski, Baranowski, Thompson and Greaves18). This might be due to the different methods of and limitations in the assessment of physical activity applied, rather than to a true absence of effects(Reference Mendoza, Zimmerman and Christakis3, Reference Jiang, Rosenqvist, Wang, Greiner, Ma and Toschke6, Reference Campbell and Hesketh9, Reference Marshall, Biddle, Gorely, Cameron and Murdey19, Reference Connelly, Duaso and Butler20). In contrast, a significant, at least moderate association of high physical activity as defined by the proposed classification and childhood overweight/obesity was observed in the present study. Children with high physical activity also performed significantly better in motor testing. This seems plausible and supports the idea that the questionnaire really measures physical activity, rather than other intermediate or confounding variables finally resulting in a lower overweight prevalence.
The physical activities assessed in the questionnaire may have different energy expenditures. A score defined on the relative contribution of the different physical activities to their energy expenditure would be optimal. Since there are no data as to the energy expenditure of e.g. skating, cycling, etc. in young children, it was not possible to define a score based on equal units of energy expenditure related to the different physical activities. Therefore we chose an approach based on distribution only: to define high and low physical activity regarding specific activities we chose the median as a cut-off point.
Well-known risk factors for childhood obesity – such as TV consumption, parental education, parental BMI and non-German citizenship(Reference Viswanath and Bond21–Reference Lampert, Mensink, Romahn and Woll23) – were confirmed in our data. Adjusting for these potential confounders did not change the estimates for the association between reported physical activity and overweight/obesity or motor testing.
Remarkably, watching TV frequently was not significantly associated with the number of side-to-side jumps whereas a significant association to overweight/obesity was observed. It is hypothesized that watching a lot of TV causes overweight through three mechanisms(Reference Robinson4): (i) displacement of physical activity; (ii) increased energy consumption while watching or caused by the effects of advertising; and (iii) reduced resting metabolism. When controlling for physical activity (i), still other pathways remain. Only the first of the suggested mechanisms, lack of physical activity (i), is likely to exert an effect on motor fitness.
Given the high return rates of the informative questionnaires, selection bias is unlikely. The regions were selected in order to include urban and more rural regions. Regional differences with respect to the number of jumps were adjusted using multivariate analysis.
Since the data were assessed in a cross-sectional study, reverse causation cannot be excluded: overweight or obese children might be less physically active because of their fatness.
Because we report associations to illustrate the validity of an assessment tool, reverse causation is not an issue: regardless of the direction of the causation, a valid assessment tool should find physically less active children who – be it in consequence or as a cause of their low physical activity – are more likely to be overweight. The potential prognostic relevance of high physical activity as defined in the present study needs to be established in prospective studies.
A limitation to be noted is that the analysis did not take into account the food intake of the participating children. Assessment of dietary habits was not feasible in this large sample collected on a surveillance basis. Since the classification as physically active is based on parental reports the results could be subject to information bias, in particular if the resulting misclassification was differential between over- and normal-weight children. Assuming that parents of overweight children – in knowledge of the beneficial effects of physical activity on overweight – over-reported physical activity, the estimates of our study would be biased towards the null. Compared with other studies ascertaining physical activity by one or two questions(Reference McGarvey, Keller, Forrester, Williams, Seward and Suttle11, Reference Fitzgibbon, Stolley, Schiffer, Van Horn, KauferChristoffel and Dyer12) (intervention studies with n 186 and 400, respectively) or outdoor playing time as a surrogate marker(Reference Burdette, Whitaker and Daniels13) (validation study, n 250), the set of questions used in the present study is conceptually different by listing children’s activities, which might be a useful complementary approach.
The current results can be seen as a successful attempt to demonstrate survey-based physical activity rating being correlated with weight status and a motor test in a large sample. Although the results are promising, validation against other methods is required to recommend this set of questions as an assessment tool in epidemiological research. This is planned in collaboration with an ongoing study involving accelerometry, pedometry, activity diary-keeping and electrocardiography monitoring.
A strength of the proposed approach to assess physical activity in pre-school children is its simplicity and high acceptance by parents: in 88·2 % of the completed questionnaires there were sufficient answers to compile the variable ‘physically active’. Other ways to measure regular physical activity require considerable effort (e.g. direct observation of physical activity)(Reference Goran24) or invasive methods (e.g. VO2max used to assess physical fitness as a proxy)(Reference Nakagaichi, Lee and Tanaka25). This simple approach therefore might be a useful tool to assess physical activity in pre-school children in studies on causes of childhood obesity. Further validation against other methods is warranted.
Acknowledgements
The GME-Study Group comprises members from: the Bavarian Health and Food Safety Authority, Oberschleißheim (Gabriele Bolte, Hermann Fromme, Annette Heißenhuber, Michael Mosetter, Gabriele Morlock, Christine Mitschek, Uta Nennstiel-Ratzel, Manfred Wildner); the Health Authority of the District Office of Bamberg (Winfried Strauch, Wiltrud Doerk, Rosemarie Sittig, Heidi Thamm); the Health Authority of the District Office of Guenzburg (Roland Schmid, Tatjana Friess-Hesse, Gudrun Winter); the Health Authority of the City Ingolstadt (Elisabeth Schneider, Traudl Tontsch, Christine Gampenrieder, Margot Motzet, Gerlinde Woelk); the Department of Health and Environment, City of Munich (Heidi Mayrhofer, Sylvia Kranebitter, Gertraud Rohrhirsch, Brigitte Weise); the Health Authority of the District Office of Schwandorf (Maximilian Kuehnel, Kornelia Baranek, Gitte Koch-Singer); the Institute of Social Paediatrics and Adolescent Medicine, Ludwig-Maximilians University, Munich (Ruediger von Kries, Ladan Baghi); and the Bavarian State Ministry of the Environment, Public Health and Consumer Protection (Bernhard Liebl).
Funding: O.B. was supported by LMUinnovativ research priority project MCHealth (sub-project I). Data management was supported by the Bavarian State Ministry of the Environment, Public Health and Consumer Protection.
Conflicts of interest: none declared by any of the authors.
Contributorship: O.B. – statistical analysis, conception and writing of the manuscript; G.B. – coordination and data management of the GME health survey, review of the manuscript; G.M. – development of questionnaire items, review of the manuscript; S.R. – consulting on the statistical analysis, review of the manuscript; R.v.K. – conceiving the evaluation concept, questionnaire items, writing the manuscript.
