Study design
The present study used data obtained in the ChecKid study 2012. ChecKid is a repeated cross-sectional study of primary school children aged 4 to 12 years from the city of Zwolle in the Netherlands. ChecKid measures were collected in 2006, 2009, and 2012. The objectives of ChecKid are to investigate trends in overweight and to examine life style behaviours related to childhood overweight and obesity and determinants of these behaviours within families, schools and neighborhoods. ChecKid is part of an integrated approach in which quantitative and qualitative monitoring research and environmental scans support the development, implementation and evaluation of tailored community wide interventions.
From all parents who participated in the study, and from those children who were 12 years and older at the time of data collection, passive consent was obtained. In the Netherlands, for children aged 12 – 16 years, consent from parents as well as consent from children themselves is required for participation in research studies. For children younger than 12 years old parental consent only is sufficient for participation, and for children older than 16 years old no parental consent is needed for participation. Medical ethical approval was obtained from the Medical Ethics Committee of the VU University Medical Centre.
Study population
A total of 43 primary schools in the city of Zwolle were invited to participate, of which 35 (81%) schools participated. When schools did not want to participate, it was mostly because of other priorities. Participating schools were equally spread over all neighborhoods in Zwolle. When a school agreed to be included in the study, all children attending the school (4-12 years) and their parents were invited to participate by means of letters distributed via the schools. For this study passive consent, which involved distributing a letter to the children’s parents and to children that were 12 years or older describing the study and instructing them to respond only if they did not want (their child) to participate, was required from all parents and from those children that were 12 years old and older. In these letters we included information on the consent procedure, and underlined the possibility for children (and parents) to end participation in the study at any time, even when consent has been obtained. There were 135 (2.2%) parents that did not give consent for participation and 34 (0.6%) children that refused to participate in the anthropometrical measurements. Children without anthropometrical data were excluded from this study because weight status of the child is a crucial variable in this study. Further exclusion criteria for participants were not being proficient in the Dutch language, being older than 12 years of age, and not living in the city of Zwolle.
A total of 3328 children aged 9 to 12 years old from the 6th, 7th and 8th grade completed questionnaires, but 1330 of those children did not have a matched parent report of health behaviours, leaving a sample of 1998 matched parent and child reports. Additionally, anthropometrical measurements of the children (height, weight and waist circumference) were performed. Eligible children included those who had an anthropometric measurement (height and weight), whose parents filled in a self-report questionnaire about the reported behaviours, and who completed a questionnaire themselves. A total of 1998 children aged 9 to 12 years old met these criteria and were included in this study.
Measurements
Anthropometric measurements
Anthropometric measurements were performed during 3 weeks in October and November 2012. Trained students measured body height, weight and waist circumference using a standardized protocol [25, 26]. Height was measured to the nearest 0.1 cm with a stadiometer, and weight was measured to the nearest 0.1 kg with a Seca digital scale. During the measurements, the children wore gym clothing and no shoes. Body Mass Index (BMI) was calculated as weight in kilograms divided by height in meters squared. The children’s age- and sex-specific BMI cut-off points suggested by Cole et al. were used to define thinness, healthy weight, overweight and obesity [27, 28]. We used the term thinness which WHO uses to mean low BMI in adults and adolescents [29]. The international BMI cut offs for child overweight and obesity are based on the adult cut offs of 25 and 30 at 18 years and cover the age range 2-18 years [28]. It would be logical to produce BMI cut offs for underweight or thinness using the same principle. However, presently, no expert guidelines for thinness exist, and the current cut-offs classifying thinness are merely based on supposition [27, 30, 31]. In addition, underweight or thinness does not have the same meaning in adults and children. In adults, underweight or thinness indicates low BMI, and can have serious health consequences and comorbidities, whereas in children underweight is low weight for age and wasting is low weight for height [29]. Cole et al. suggest extending the adult term of thinness to children, meaning low BMI for age [27]. For these reasons, and because the prevalence rates of thinness (9%) and obesity (1.4%) in our study were relatively low, we grouped children who were not overweight and defined them as ‘healthy-weight children’ and grouped children who were overweight and obese and defined them as ‘overweight’.
Questionnaires
The ChecKid children’s questionnaire consisted of questions on health-related lifestyle behaviours (diet, physical activity, sleeping habits, sedentary behaviour) and determinants of these behaviours (e.g. home and school environments) and was designed for children aged 9 -12 years of age attending grade 6, 7 and 8 in Dutch primary schools. The children’s questionnaire concerned children’s behaviour during a regular schoolday as we were especially interested in finding indications for interventions that could possibly be implemented or supported in a school setting. The ChecKid parental questionnaire consisted of questions on the same subjects but also included socio-demographic variables such as the child’s age, gender, postal code, ethnicity (assessed by country of birth of both parents) and socio-economic status (SES) (assessed by educational level of parents). Existing validated questionnaires on health behaviour were used for the design of the questionnaires [32, 33].
Because a limited amount of questions was worded in exactly the same way in both the parent and children’s questionnaires, we could only use these questions for our analyses on the level of agreement between parent and child reports. Questions worded in exactly the same way on the parent and child questionnaires were used for the analyses regarding parent-child agreement. For example, we asked children the following questions ‘On how many days do you eat breakfast before going to school during the schoolweek?’ and ‘On how many days do you eat dinner at the dining table with your parents during the schoolweek?’. The corresponding questions for the parents were ‘On how many days does your child eat breakfast before going to school during the schoolweek’ and ‘On how many days do you and your child eat together at the dining table during the schoolweek?’.Children could respond with: (almost) never; 1 day per week; 2 days per week; 3 days per week; 4 days per week; 5 days per week, and the corresponding response categories for parents were: 0 or < 1; 1; 2; 3; 4; 5 days in a regular school week. For the exact questions used see Additional file 1.
Health behaviours
We investigated the level of agreement between parent and children reports with respect to five important health related behaviours: breakfast consumption; family dinner; outside play; means of transportation to school and TV/DVD viewing. Outside play was used as indicator of the child’s physical activity, and TV/DVD viewing was used as an important indicator of sedentary behaviour. Family dinner and breakfast consumption were used as indicators of the child’s dietary behaviour. As the main purpose of this study was to examine agreement between the reports, we only used five questions which were worded identically in the parental and children’s questionnaires. We were aware that that the examined behaviours were used as indicators of the specific behaviours, and thus may not represent the wider health related behaviour.
Frequency of breakfast consumption on schooldays and frequency of eating a family dinner together at the table on schooldays in both parents and children were used as indicators of meal patterns.
Outside play was used as an indicator for physical activity and was measured by investigating time spent on outside play. Parents and children were asked to report frequency and duration of time (in categories) spent on outside play. Average time per day spent on the behaviour was calculated by multiplying the number of days that the child spent on the behaviour by the mid-category values of duration of the item in 5 categories: < 0.5, 0.5-1, 1-2, 2-3, and > 3 h a day, and dividing this by 5; the number of schooldays per week. The categories ‘2-3 h’ and ‘more than 3 hours’ were combined so that the response categories in the parents’ reports were the same as they were for the children’s reports. Current recommendations for children aged 5 to 17 years are to spend at least 60 min per day on outside play [34]. Therefore, outdoor play was dichotomized as < 60 and ≥ 60 min per day.
TV/DVD viewing was used as an indicator for sedentary behaviour, as TV viewing has been known to be an important determinant for the development of overweight [35]. Parents and children were asked to report frequency and duration of time (in categories) spent watching TV/DVD. Average time per day spent on the behaviour was calculated by multiplying the number of days that the child spent on the behaviour by the mid-category values of duration of the item in 5 categories: < 0.5, 0.5-1, 1-2, 2-3, and > 3 h a day, and dividing this by 5; the number of schooldays per week. The categories ‘2-3 h’ and ‘more than 3 hours’ were combined so that the response categories in the parents’ reports were the same as they were for the children’s reports. Current recommendations for children aged 4 to 17 years are not to use screentime for more than 2 h per day [36, 37]. Thus, TV/DVD viewing was dichotomized as < 2 and ≥ 2 h per day.
Means of transportation to school could be indicated by the following options: cycling; walking; on the back of a scooter; on the back of a bicycle; brought by car; by bus; other.
Statistical analyses
Statistical analyses were conducted using the PASW 20.0 and Stata 11 (StataCorp, College Station, Texas) software packages. Descriptive statistics were used (mean, standard deviations and percentages) to describe the study sample and the differences in parent reports of the behaviours and child reports of the behaviours.
Level of agreement between parent proxy reports and child-self reports.
To assess the level of agreement between child and parent reports about frequency of breakfast consumption, frequency of family dinner, average duration and frequency of outside play and average duration and frequency of TV/DVD viewing, we compared calculated averages of frequency and duration of the studied behaviours. To do so, the weighted kappa statistic was used. The response categories of these variables are ordinal which means that not every disagreement can be weighted the same; for example, a difference between categories of ‘0 days per week’ and ‘5 days per week’ is a more serious discrepancy than a difference between categories of ‘3 days per week’ and ‘4 days per week’. In this study, we used the non-weighted kappa statistic to determine the level of agreement between child and parent reports on the means of transportation to school, because of the categorical response categories. The non-weighted kappa statistic does not take the extent of disagreement in account, every disagreement is weighted evenly [38, 39]. To classify the strength of agreements the standards of Landis and Koch were used for the kappa coefficients: ≤0 = poor, 0.01–0.20 = slight, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = substantial, and 0.81–1.0 = almost perfect [38].
The level of agreement between children and their parents was compared between categories of children’s weight status in stratified analyses. We calculated kappa CI’s for healthy-weight and overweight children and compared these (Table 3).
Level of agreement between parent reports and child reports in healthy-weight and overweight children.
We also explored whether children reported more, less or the same amount of the health related behaviour as their parents. The parent-child dyads were categorized into three categories: 1) children reporting the same frequency or duration of the health related behaviour as their parent; 2) children reporting lower frequency or shorter duration of the health related behaviour than their parent (i.e., less hours or days per week of TV/DVD viewing or outdoor play, or less days on which they ate breakfast and participated in a family dinner); 3) children reporting higher frequency or longer duration of the health related behaviour than their parent (i.e., children reported more hours or days per week of TV/DVD viewing or outdoor play than parents, or more days on which they ate breakfast and participated in a family dinner). Children’s weight status and the reporting categories were explored using multinomial logistic regression analysis. First, crude analyses were performed. Second, adjusted analyses were carried out, controlling for potential confounding effects of gender, SES and ethnicity, weight status and age of the parent.