Objective: To determine trends in the prevalence of overweight and obesity in children in The Hague (the Netherlands) from 1999 through 2007.
Design: Population-based study of a series of cross-sectional assessments of height and weight from electronic health records.
Setting: Child Health Care (Municipal Health Service), The Hague.
Participants: 50 961 children aged 3–16 years, with Dutch (59%), Turkish (17%), Moroccan (13%) or Surinamese South Asian (11%) ethnicity, representative of the four major ethnic groups in The Hague, with 85 234 weight and height measurements recorded in 1999–2007.
Main outcome measures: (Trends in) the prevalence of overweight (excluding obesity) and obesity as defined by the International Obesity Taskforce cut-off points, using logistic regression with year as independent variable.
Results: From 1999 through 2007 there was a decrease in the prevalence of overweight in Dutch girls from 12.6% to 10.9% (OR 0.96; 95% CI 0.95 to 0.98) and an increase in Turkish boys from 14.6% to 21.4% (OR 1.08; 95% CI 1.04 to 1.11). Obesity prevalence rose significantly in Turkish boys from 7.9% to 13.1% (OR 1.04; 95% CI 1.01 to 1.06) and in Turkish girls from 8.0% to 10.7% (OR 1.04; 95% CI 1.01 to 1.08). Dutch boys, and Moroccan and Surinamese South Asian boys and girls showed no significant trends.
Conclusions: The declining prevalence of overweight in Dutch girls may indicate reversal of previous trends in the Netherlands. However, in Turkish children overweight prevalence and obesity is high and increasing. Further public health action is necessary, especially for Turkish children.
Statistics from Altmetric.com
Worldwide, the prevalence of obesity in children has steadily increased since the 1980s.1 2 The fourth Dutch growth study of 1997 in the Netherlands also showed a two- to fourfold increase in overweight and obesity prevalence in Dutch children compared to 1980 with an overall overweight (including obesity) prevalence of 13%.3 In this study, for the first time, the second and third largest ethnic groups in the Netherlands (the Turkish and Moroccan) were studied separately. These groups showed a two to four times higher prevalence of both overweight and obesity compared to Dutch children.4 The prevalence of overweight and obesity in Turkish children in the Netherlands is, however, comparable to that found in Germany.5 In Turkey itself, the prevalence of overweight and obesity in children is lower and approximates the current level of overweight and obesity in Dutch children in the Netherlands.6 7 8
Compared with many other Western European countries, the prevalence of overweight and obesity in Caucasian children in the Netherlands is relatively low. In 1997 approximately 12% of children aged 7–17 years were overweight or obese, while in the United Kingdom rates in Caucasian children were almost twice as high as in the Netherlands.9 However, the prevalence of overweight and obesity in the Netherlands may be changing. A recent study showed a faster rate of increase of overweight and obesity in the Netherlands since 1997 than between 1980 and 1997.10 Since in that study no separate analyses of overweight and obesity in relation to ethnicity were performed, it is not known whether the prevalence of overweight or obesity has increased in all ethnic groups at the same rate.
What is already known on this topic
Overweight and obesity prevalence in children in the Netherlands is still rising and at a faster rate than between 1980 and 1997.
Overweight and obesity in the Netherlands are more prevalent in Turkish and Moroccan children.
What this study adds
In Dutch girls overweight prevalence has declined since 1999.
In Turkish children overweight and obesity prevalence has risen strongly since 1999.
The aim of our study was to determine time trends in prevalence of overweight (excluding obesity) and obesity among children in the largest ethnic groups in the Netherlands between the ages of 3 and 16 years, using data routinely collected in the city of The Hague between 1999 and 2007.
Child Health Care in the Netherlands records growth data as part of a routine health surveillance programme covering all children. Height and weight are measured at specific ages: 3–4, 5–6, 7–10, and 13–16 years of age. In general all children in these age groups are invited for a standard preventive health examination by a school physician or school nurse. Examinations of 3–6-year olds are performed by a school physician and 7–10-year olds are examined by a school nurse. Until 2004, 13–16-year olds were examined by school physicians, but the numbers of examined adolescents were never optimal because of understaffing. From 2004 examinations of 13–16-year olds were prioritised and more school nurses were employed to perform the examinations instead of school physicians. The year 2004 can be regarded as a transition year in which more 13–16-year olds and slightly less 7–10-year olds had a health assessment.
In the city of The Hague, with a population of around 470 000, all findings from preventive health assessments by Child Health Care have been documented in an electronic patient record system since September 1998. For the current study we used all height and weight data that had been collected between 1 January 1999 and 31 December 2007. Only children with Dutch, Turkish, Moroccan or Surinamese South Asian ethnicity were included in the study. Main personal data such as name, date of birth, sex, address, postal code, land of birth of both child and parents, and nationality were acquired from the Municipal Database (Personal Files) to ensure the electronic patient record system was kept up to date.
Anthropometric methods and definitions of overweight
Heights and weights were measured by trained public health care professionals (school physicians and school nurses). At ages 3 through 6, body weight was measured with a standard mechanical or electronic step scale in underclothes. From the age of 7, body weight was measured in light clothing without shoes. Weight was rounded to the nearest 0.1–0.5 kg. Height was measured with a stadiometer or microtoise and rounded to the nearest 0.1–0.5 cm. Body mass index (BMI) was calculated with the formula: weight (in kg)/(height(in m))2.
To determine weight status we used the internationally agreed standard for overweight (excluding obesity) and obesity of the International Obesity Taskforce (IOTF) using BMI cut-off points for age and sex that correspond to the adult cut-offs of 25 for overweight and 30 for obesity.11 In this study we will define overweight as overweight but not obese, and total overweight as overweight including obesity.
Ethnicity and socioeconomic status
The ethnic origin of children was first based on the mother’s country of birth. However, if the mother was born in the Netherlands and the father was born outside the Netherlands, the father’s birth country prevailed. If the parental country of birth was not recorded in the municipal database, nationality was used as a proxy for ethnicity. Surinamese South Asian children could not be selected on parental birth country alone because Surinam is a multiethnic society with people originating from China, Indonesia, India, the Netherlands and Africa. Therefore, to select Surinamese South Asian children we matched the family names with a list of 2236 typical Surinamese South Asian family names.
As a proxy of socioeconomic status (SES) we used municipal area deprivation scores (ADS) that have been attributed to each residential district of The Hague since 1995. ADS is a continuous variable that is based on unemployment rates, average income, housing subsidy rates in the particular area, and percentage of immigrants. It ranges from −25 to 25. The ADS was added to each recorded height and weight based on the postal code at the time of measurement.
Method of analysis
Dutch, Turkish and Moroccan growth reference values from the fourth Dutch growth study3 4 were added to the individual records. As there are no growth reference values for Surinamese South Asian children, the Dutch reference was added to these records. Extreme values of height, weight, BMI, height-SDS, weight-SDS and BMI-SDS were checked and either corrected where possible or excluded from the analyses. Time trends in prevalence rates of overweight and obesity over the 9-year period (1999–2007) were calculated with logistic regression analyses with year of examination as a continuous independent variable for each sex, ethnic group and age group separately. In these analyses, all records of children in the database that did not belong to the analysed group were used as a reference group, that is all non-obese (including overweight) were analysed versus obese and all non-overweight (including obese) versus overweight. Age and SES, both measured on a continuous scale, were introduced into the model as adjusting variables. p Values (two-sided) less than 0.05 were considered statistically significant.
A total of 85 234 measurements of both weight and height in 50 961 children, measured between 1999 and 2007, were used in the analyses. Less than 0.1% of all measurements were excluded from the analyses. Table 1 shows the details of the studied population. Ethnic groups are represented equally over the years and together form a representative sample of the population of The Hague.
Participation in the health surveillance programme was high (83.0%) after non-respondents were sent a second invitation to attend. Participation did not change markedly during the studied period and the participation rate was similar between ethnic groups (Dutch 83.5%, Turkish 82.0%, Moroccan 81.6%, Surinamese South Asian 83.0%). Reasons for non-participation are not known.
Over the 9-year period, Turkish children had the highest mean prevalence of total overweight (including obesity) of 28%, followed by Moroccan (23%), Surinamese South Asian (15%) and Dutch children (13%). In Turkish boys (fig 1) and girls (fig 2) total overweight prevalence increased from 22.4% to 34.5% (p<0.001) and from 27.4% to 33.8% (p = 0.003), respectively. Total overweight in Dutch girls decreased significantly from 16.3% in 1999 to 14.2% in 2007 (p<0.001). Dutch boys, and Moroccan and Surinamese South Asian boys and girls show no significant trend after adjustment for age and SES.
In tables 2 and 3, the prevalence of overweight and obesity, respectively, is shown by ethnic group, sex and age group. The odds ratios are adjusted for SES (for the different age groups) or age and SES (total per sex).
The analyses show a decline in the prevalence of overweight in Dutch 3–6-year-old boys and girls; however, overweight prevalence in 7–10-year-old Dutch boys showed a significant increase and therefore prevalence in the group of Dutch boys as a whole seems to be stable.
In Turkish boys overweight prevalence increased from 14.5% in 1999 to 21.4% in 2007 (p = 0.005), which can be attributed to a significant increase in overweight in 7–10-year-old boys. Obesity prevalence, however, increased in all age groups and the total prevalence rose from 7.9% to 13.1% in 2007 (p<0.001). In Turkish girls overweight increased significantly in 3–6-year olds and obesity in 7–10-year olds. While at a group level there were no significant trends for overweight and obesity in Moroccan and Surinamese South Asian children, overweight prevalence significantly increased in 7–10-year-old Surinamese South Asian boys and 13–16-year-old Moroccan boys. In Moroccan 13–16-year-old boys obesity rates also rose.
The main finding of this study is that in recent years overweight and obesity prevalence in children in the city of The Hague (the Netherlands) has been rising, particularly among Turkish children. The prevalence of total overweight in Turkish boys rose from 22.4% in 1999 to 34.5% in 2007, which is an increase of 54% since 1999. In Turkish girls there was a 23% increase in total overweight since 1999, rising from 27.4% to 33.8%. We also found an unexpected decline in the prevalence of overweight (excluding obesity) in Dutch girls from 12.6% in 1999 to 10.9% in 2007, a 14% decrease since 1999. These findings differ from a previous study that showed an increase in overweight and obesity prevalence among children in the general Dutch population.10 However, in that study ethnic differences were not investigated. Thus, ethnic specific figures in that study may have been similar to those in our study. Alternatively, differences between the studies may be a result of the population of children in The Hague perhaps differing from other regions in the Netherlands or public health campaigns potentially having been more successful in The Hague. More national growth studies are needed to confirm if our findings are representative of the entire Netherlands.
The use of growth data from standard health assessments is ideal for monitoring overweight, and the increasing use of electronic patient record systems makes it much easier to acquire growth data for analyses. A strength of this study is the large study population, high response rate and availability of measured (rather than reported) data on weight and height. Furthermore, the reliability of the collected data is high. The electronic patient record system in which the growth data are recorded also plots growth charts which are shown to parents and children during the health examination. Errors in registration are therefore easily recognised and are corrected by the examiner. Another strong point of our study is that data on SES (based on ADS) were available. A limitation of using ADS as a proxy for SES is that it is partly based on the percentage of immigrants in the particular area. When used as an adjusting factor it could lead to overcompensation for ethnicity. In the present study SES was only used as adjusting variable in analyses within the different ethnic groups and therefore did not affect the outcome.
Although response rates were generally high, the entire target population of 13–16-year olds was not invited every year for a health assessment because of understaffing until 2003. This would only have affected the results if a specific group had been invited to attend the health assessments. Since this was not the case, the examined children were most likely a random sample of the target population.
The decrease in the proportion of overweight (excluding obese) Dutch girls together with no overall increase in overweight and obesity prevalence in Dutch boys may suggest that in Dutch children a plateau in overweight and obesity prevalence has been reached. Similarly, in Sweden a decrease of overweight in 10-year old girls,12 and in France a stabilisation of overweight (including obesity) in 4–15-year-old children have recently been found.13
It is unclear why in our study the prevalence of overweight and obesity is increasing in Turkish children, while overweight has decreased in Dutch 3–6-year-old boys and girls. Perhaps public health campaigns of the past decade have been better directed to and adopted by Dutch parents and children than their Turkish counterparts. A low parental educational level, a known risk factor for childhood overweight and obesity,3 may be an important influential factor since Turkish people in the Netherlands belong to the lowest educated ethnic groups.14
Many social and behavioural factors such as parenting skills, self-regulation of food intake and the parents’ own eating behaviour influence the development of overweight and obesity.15 The decrease of overweight (excluding obesity) in the youngest age group of Dutch children could mean that parents of young children are becoming more aware of the importance of physical activity and healthy nutrition, and have changed some of these behaviours. However, the concurrent rise in overweight prevalence in 7–10-year-old Dutch boys shows that explaining these trends may be more complex with many factors influencing the weight status of a population.
In conclusion, the prevalence of overweight and obesity in the Netherlands may have reached a plateau in most ethnic groups. However, in Turkish children overweight as well as obesity is still increasing at an disturbing rate. In addition, since prevalence in all ethnic groups continues to be much higher than in 1980, public health interventions are still needed, especially interventions tailored to Turkish children and their parents.
We would like to thank Ron Smit for his help with extracting the data from the electronic patient record system.
Funding This study was supported by the Municipality of The Hague, Leiden University Medical Center and TNO Quality of Life.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.