Acta Pædiatrica ISSN 0803-5253
REGULAR ARTICLE
Poor school performance is associated with a larger gain in body mass index during puberty M Kark ([email protected])1, A Hjern2,3, F Rasmussen1,4 1.Unit of Child and Adolescent Health, Centre for Epidemiology and Community Medicine, Health Care Services, Stockholm County Council, Stockholm, Sweden 2.Centre for Health Equity Studies (CHESS), Karolinska Institutet/Stockholm University, Stockholm, Sweden 3.Clinical Epidemiology, Department of Medicine, Karolinska Institutet, Stockholm, Sweden 4.Child and Adolescent Public Health Epidemiology Group, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden
Keywords Body mass index, COMPASS study, Growth, School performance, Type 2 diabetes Correspondence M Kark, Centre for Epidemiology and Community Health, Box 1497 171 29 Solna, Sweden. Tel:+46 8 457 23 42 | Email: [email protected]. Received 2 April 2013; revised 3 September 2013; accepted 14 October 2013. DOI:10.1111/apa.12471
ABSTRACT Aim: Social inequalities in type 2 diabetes and coronary heart disease may be established in formative school years. We investigated whether school performance is associated with adiposity and increase in body mass index (BMI) between 10 and 15 years of age. Methods: A community sample of 2633 school children had height and weight measured in school at the ages of 10 and 15. Percentages of body fat and waist circumference were measured at the age of 15. Mean grades in several school subjects at the age of 15 (ninth school year) were divided into quartiles. A linear regression analysis with BMI as the main outcome took into account parental education and ethnicity, obtained from registers, and children’s living habits, collected by questionnaires. Results: In adjusted models, longitudinal changes in BMI between the ages of 10 and 15 were larger in the lowest quartiles of school grades compared with the highest: for girls, they were b = 0.45 (p = 0.007) and for boys they were b = 0.45 (p = 0.016). Crosssectional regression analyses, with percentage of body fat and waist circumference as outcomes, showed similar results. Conclusion: Our results suggest that school performance is one pathway to social inequalities in obesity in school children.
BACKGROUND Socio-economic differences in coronary heart disease and type 2 diabetes in adulthood have been described in numerous studies (1,2). In the Nordic welfare states, these social differentials are obvious for educational level and occupational socio-economic position (3). Socio-economic position in childhood is also a strong risk factor for coronary heart disease and type 2 diabetes (4,5), and it is conceivable that these associations are partially mediated by achievements or failure in the school system. Social inequalities in these chronic diseases are representations of societal hierarchies (2). The work environment plays an increasingly important role in development of coronary heart disease and type 2 diabetes (6). A recent review confirmed that work-related psychosocial factors, such as high psychological demands combined with lack of control and social support, are associated with higher risks of coronary heart disease (7) and type 2 diabetes (8). Structural factors, such as national wealth, income inequality and access to education, are determinants of adolescent health (9), and obesity in adolescence is strongly related to childhood socio-economic position (10). The school can be seen as the children’s work environment, where students with emotional problems, impaired learning ability or language difficulties may perceive demands that
they cannot meet and may feel that they lack control and social support. It is known from psychology that promotion of self-regulation, that is, a balance between processes of emotional–motivational arousal and cognitive control processes, can foster school readiness and prevent school failure (11). There is also some evidence that suboptimal self-regulatory skills predict development of obesity in preschool age (12) and that chronic stress in childhood can diminish children’s self-regulatory capacity (13). Evans recently reported increased weight gain between nine and 11 years of age in children exposed to increased social
Key notes
©2013 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd 2014 103, pp. 207–213
Adolescence is apparently a key period for weight gain among socially disadvantaged children, and this higher weight will often persist into adulthood. Our study of 2633 children showed that longitudinal changes in BMI between the ages of 10 and 15 were larger in the lowest quartiles of school grades than the highest. The findings suggest that school performance is a possible pathway that can lead to social stratification of obesity.
207
Poor school performance and BMI
Kark et al.
disadvantage and psychosocial stress and provided some evidence that the longitudinal association was at least partially mediated by low self-regulatory capacity (14). Hierarchies are formed among children in schools according to achievements. The most obvious indicators of school success are the school grades awarded at the end of the school year. Childhood socio-economic position is associated with school performance, which is an important early indicator of a social career path, later defining the social status in adulthood. School performance is also associated with poor health behaviours, which persist into adulthood, such as smoking (15), alcohol habits (16) and obesity (17). Smoking and increased BMI and waist circumference in adolescence occur often in constellation with elevated blood pressure, low grade inflammation and insulin resistance, and all these risk factors track into adulthood and are predictive of coronary heart disease and type 2 diabetes in later life (18). A recent epidemiological study from Denmark found that weight gain in men from young adulthood to early middle age increased all-cause mortality regardless of initial weight (19). Another recent study from Japan reported that long-term weight gain from early adulthood to middle age increased the risk of type 2 diabetes in both genders (20). In Sweden, the educational policy for several decades has been, as much as possible, to refrain from creating social hierarchies in classrooms by not awarding grades until the last 2 years of primary school. Average grades from the final ninth year of compulsory school in Sweden summarise several years of school work and are given much attention as they are used as selection criteria for secondary school. The aim of this study was to investigate whether school performance (a summary score of the average grades from the final ninth year of compulsory school) is predictive of degree of adiposity and increase in BMI between 10 and 15 years of age, as possible markers of long-term risks of coronary heart disease and type 2 diabetes. It is possible that school performance constitutes a key link between childhood socio-economic position, increased risk of obesity in adolescence and social differentials in coronary heart disease and type 2 diabetes in adulthood.
METHODS Study population The COMPASS study, founded in the year 2000, is a cohort study of children from a geographically well-defined area of Stockholm County with approximately 370 000 inhabitants in 2001. All 4188 children, born in Sweden or abroad, between 1985 and 1987, attending the eighth grade of municipal schools in 2000 or ninth grade in 2001, were identified and invited to participate in the study which included a health examination at approximately 15 years of age. Parents received an information letter, mailed to their home, in which they were requested to respond to the researchers if their child were not allowed to participate in the study or did not want to participate. If such a letter was not received, the child was approached in school and asked
208
to participate. The children could also decline to participate when they met the research staff at school without giving any reason. Of these adolescents, 10.5% declined to participate and 4.8% did not participate because of sick leave, a severe chronic disorder or unexplained absence from school at the time of data collection. In total, 3548 of all 4,188 (84.7%) children participated in the data collection between October 2000 and April 2002 as described elsewhere (21,22). Measured anthropometric data were also collected from school records as described below. This study includes 2633 children, 1348 boys and 1285 girls, who had complete data on parental education, ethnicity, height and weight measurement at 10 and 15 years of age, recorded age at the health examination and self-reported information about smoking habits, breakfast habits and physical activity. The proportion of young people with immigrant background was 41%, which is considerably higher than in Stockholm County as a whole (22% of children living at home aged 0–17, in 2005, according to Statistics Sweden). The study was approved by the Ethical Review Board in Stockholm, Sweden. Study variables Data from the following sources were used in this study: (i) anthropometric data from routine child health records (ii) anthropometric data collected by project nurses at the age of 15 years (iii) self-reported information from the COMPASS cohort members by questionnaires and (iv) sociodemographic data and school information from national registers held by Statistics Sweden linked to the study population through the children’s personal identification (ID) number and to their biological parents through the Statistics Sweden’s Multi-Generation Register. Information on height and weight at 10 years of age was collected from school health records. Project staff nurses measured weight, height and waist circumference at schools in the eighth or ninth school years. Height was measured twice to the nearest 0.5 cm and weight to the nearest 100 g. The mean of the two measurements was used to calculate BMI at 15 years of age. Bioelectric impedance was measured using Quantum II, RJL Systems, Clinton Township, MI, USA, as described elsewhere (21,22), and fat-free mass was calculated using the cross-validated equation by Houtkooper et al. (23). Fat mass was then calculated by subtracting fatfree mass from total body mass. The percentage of fat mass was calculated as fat mass divided by total body mass in percentage. Waist circumference was also measured twice to the nearest 1 mm at the minimum circumference between the iliac crest and the rib cage, and the average of the two measurements was used. Differences in BMI between 10 and 15 years of age were calculated by subtracting the BMI value at 15 years of age with the BMI value at 10 years of age. School grades from the final ninth year of compulsory school were obtained from the National School Register when the study participants were 15–16 years old. The reported mean value was based on the student’s 16 best subjects. Each subject provides 10–20 points, and by summing up all 16 subjects, the maximum number of points is 320
©2013 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd 2014 103, pp. 207–213
Kark et al.
points and the minimum is 10 points. This mean value was used to create four categories (quartiles), with boys and girls taken together. The mean value in the first quartile was 128.1 points (range: 10–170), the mean value in the second quartile was 189.4 (175–205), the third quartile was 227.4 (210–245), and the fourth quartile was 276.6 (250–320). The Swedish National School Register provides high-quality data, and summary statistics are published regularly (24). Information on maternal and paternal educational level was obtained from the Statistics Sweden’s Register of Education in 2000 when the cohort members were of preschool age. The highest level attained by the mother or the father was used in this study and categorised as low (9 years or less), medium (secondary school) and high (postsecondary education). Parental country regions of birth were categorised as Sweden, west Europe, east Europe, outside Europe and mixed. The mixed category includes parents from diverse country regions of birth. Data collected by questionnaires, as described above, were used to create the variables about daily smoking (yes/no) and frequency of eating breakfast during a normal school week (0–1 days/2–3 days/4–5 days per week). Physical activity levels were categorised as low, medium and high active. Low active was defined as moderate + hard physical activity
REGULAR ARTICLE
Poor school performance is associated with a larger gain in body mass index during puberty M Kark ([email protected])1, A Hjern2,3, F Rasmussen1,4 1.Unit of Child and Adolescent Health, Centre for Epidemiology and Community Medicine, Health Care Services, Stockholm County Council, Stockholm, Sweden 2.Centre for Health Equity Studies (CHESS), Karolinska Institutet/Stockholm University, Stockholm, Sweden 3.Clinical Epidemiology, Department of Medicine, Karolinska Institutet, Stockholm, Sweden 4.Child and Adolescent Public Health Epidemiology Group, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden
Keywords Body mass index, COMPASS study, Growth, School performance, Type 2 diabetes Correspondence M Kark, Centre for Epidemiology and Community Health, Box 1497 171 29 Solna, Sweden. Tel:+46 8 457 23 42 | Email: [email protected]. Received 2 April 2013; revised 3 September 2013; accepted 14 October 2013. DOI:10.1111/apa.12471
ABSTRACT Aim: Social inequalities in type 2 diabetes and coronary heart disease may be established in formative school years. We investigated whether school performance is associated with adiposity and increase in body mass index (BMI) between 10 and 15 years of age. Methods: A community sample of 2633 school children had height and weight measured in school at the ages of 10 and 15. Percentages of body fat and waist circumference were measured at the age of 15. Mean grades in several school subjects at the age of 15 (ninth school year) were divided into quartiles. A linear regression analysis with BMI as the main outcome took into account parental education and ethnicity, obtained from registers, and children’s living habits, collected by questionnaires. Results: In adjusted models, longitudinal changes in BMI between the ages of 10 and 15 were larger in the lowest quartiles of school grades compared with the highest: for girls, they were b = 0.45 (p = 0.007) and for boys they were b = 0.45 (p = 0.016). Crosssectional regression analyses, with percentage of body fat and waist circumference as outcomes, showed similar results. Conclusion: Our results suggest that school performance is one pathway to social inequalities in obesity in school children.
BACKGROUND Socio-economic differences in coronary heart disease and type 2 diabetes in adulthood have been described in numerous studies (1,2). In the Nordic welfare states, these social differentials are obvious for educational level and occupational socio-economic position (3). Socio-economic position in childhood is also a strong risk factor for coronary heart disease and type 2 diabetes (4,5), and it is conceivable that these associations are partially mediated by achievements or failure in the school system. Social inequalities in these chronic diseases are representations of societal hierarchies (2). The work environment plays an increasingly important role in development of coronary heart disease and type 2 diabetes (6). A recent review confirmed that work-related psychosocial factors, such as high psychological demands combined with lack of control and social support, are associated with higher risks of coronary heart disease (7) and type 2 diabetes (8). Structural factors, such as national wealth, income inequality and access to education, are determinants of adolescent health (9), and obesity in adolescence is strongly related to childhood socio-economic position (10). The school can be seen as the children’s work environment, where students with emotional problems, impaired learning ability or language difficulties may perceive demands that
they cannot meet and may feel that they lack control and social support. It is known from psychology that promotion of self-regulation, that is, a balance between processes of emotional–motivational arousal and cognitive control processes, can foster school readiness and prevent school failure (11). There is also some evidence that suboptimal self-regulatory skills predict development of obesity in preschool age (12) and that chronic stress in childhood can diminish children’s self-regulatory capacity (13). Evans recently reported increased weight gain between nine and 11 years of age in children exposed to increased social
Key notes
©2013 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd 2014 103, pp. 207–213
Adolescence is apparently a key period for weight gain among socially disadvantaged children, and this higher weight will often persist into adulthood. Our study of 2633 children showed that longitudinal changes in BMI between the ages of 10 and 15 were larger in the lowest quartiles of school grades than the highest. The findings suggest that school performance is a possible pathway that can lead to social stratification of obesity.
207
Poor school performance and BMI
Kark et al.
disadvantage and psychosocial stress and provided some evidence that the longitudinal association was at least partially mediated by low self-regulatory capacity (14). Hierarchies are formed among children in schools according to achievements. The most obvious indicators of school success are the school grades awarded at the end of the school year. Childhood socio-economic position is associated with school performance, which is an important early indicator of a social career path, later defining the social status in adulthood. School performance is also associated with poor health behaviours, which persist into adulthood, such as smoking (15), alcohol habits (16) and obesity (17). Smoking and increased BMI and waist circumference in adolescence occur often in constellation with elevated blood pressure, low grade inflammation and insulin resistance, and all these risk factors track into adulthood and are predictive of coronary heart disease and type 2 diabetes in later life (18). A recent epidemiological study from Denmark found that weight gain in men from young adulthood to early middle age increased all-cause mortality regardless of initial weight (19). Another recent study from Japan reported that long-term weight gain from early adulthood to middle age increased the risk of type 2 diabetes in both genders (20). In Sweden, the educational policy for several decades has been, as much as possible, to refrain from creating social hierarchies in classrooms by not awarding grades until the last 2 years of primary school. Average grades from the final ninth year of compulsory school in Sweden summarise several years of school work and are given much attention as they are used as selection criteria for secondary school. The aim of this study was to investigate whether school performance (a summary score of the average grades from the final ninth year of compulsory school) is predictive of degree of adiposity and increase in BMI between 10 and 15 years of age, as possible markers of long-term risks of coronary heart disease and type 2 diabetes. It is possible that school performance constitutes a key link between childhood socio-economic position, increased risk of obesity in adolescence and social differentials in coronary heart disease and type 2 diabetes in adulthood.
METHODS Study population The COMPASS study, founded in the year 2000, is a cohort study of children from a geographically well-defined area of Stockholm County with approximately 370 000 inhabitants in 2001. All 4188 children, born in Sweden or abroad, between 1985 and 1987, attending the eighth grade of municipal schools in 2000 or ninth grade in 2001, were identified and invited to participate in the study which included a health examination at approximately 15 years of age. Parents received an information letter, mailed to their home, in which they were requested to respond to the researchers if their child were not allowed to participate in the study or did not want to participate. If such a letter was not received, the child was approached in school and asked
208
to participate. The children could also decline to participate when they met the research staff at school without giving any reason. Of these adolescents, 10.5% declined to participate and 4.8% did not participate because of sick leave, a severe chronic disorder or unexplained absence from school at the time of data collection. In total, 3548 of all 4,188 (84.7%) children participated in the data collection between October 2000 and April 2002 as described elsewhere (21,22). Measured anthropometric data were also collected from school records as described below. This study includes 2633 children, 1348 boys and 1285 girls, who had complete data on parental education, ethnicity, height and weight measurement at 10 and 15 years of age, recorded age at the health examination and self-reported information about smoking habits, breakfast habits and physical activity. The proportion of young people with immigrant background was 41%, which is considerably higher than in Stockholm County as a whole (22% of children living at home aged 0–17, in 2005, according to Statistics Sweden). The study was approved by the Ethical Review Board in Stockholm, Sweden. Study variables Data from the following sources were used in this study: (i) anthropometric data from routine child health records (ii) anthropometric data collected by project nurses at the age of 15 years (iii) self-reported information from the COMPASS cohort members by questionnaires and (iv) sociodemographic data and school information from national registers held by Statistics Sweden linked to the study population through the children’s personal identification (ID) number and to their biological parents through the Statistics Sweden’s Multi-Generation Register. Information on height and weight at 10 years of age was collected from school health records. Project staff nurses measured weight, height and waist circumference at schools in the eighth or ninth school years. Height was measured twice to the nearest 0.5 cm and weight to the nearest 100 g. The mean of the two measurements was used to calculate BMI at 15 years of age. Bioelectric impedance was measured using Quantum II, RJL Systems, Clinton Township, MI, USA, as described elsewhere (21,22), and fat-free mass was calculated using the cross-validated equation by Houtkooper et al. (23). Fat mass was then calculated by subtracting fatfree mass from total body mass. The percentage of fat mass was calculated as fat mass divided by total body mass in percentage. Waist circumference was also measured twice to the nearest 1 mm at the minimum circumference between the iliac crest and the rib cage, and the average of the two measurements was used. Differences in BMI between 10 and 15 years of age were calculated by subtracting the BMI value at 15 years of age with the BMI value at 10 years of age. School grades from the final ninth year of compulsory school were obtained from the National School Register when the study participants were 15–16 years old. The reported mean value was based on the student’s 16 best subjects. Each subject provides 10–20 points, and by summing up all 16 subjects, the maximum number of points is 320
©2013 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd 2014 103, pp. 207–213
Kark et al.
points and the minimum is 10 points. This mean value was used to create four categories (quartiles), with boys and girls taken together. The mean value in the first quartile was 128.1 points (range: 10–170), the mean value in the second quartile was 189.4 (175–205), the third quartile was 227.4 (210–245), and the fourth quartile was 276.6 (250–320). The Swedish National School Register provides high-quality data, and summary statistics are published regularly (24). Information on maternal and paternal educational level was obtained from the Statistics Sweden’s Register of Education in 2000 when the cohort members were of preschool age. The highest level attained by the mother or the father was used in this study and categorised as low (9 years or less), medium (secondary school) and high (postsecondary education). Parental country regions of birth were categorised as Sweden, west Europe, east Europe, outside Europe and mixed. The mixed category includes parents from diverse country regions of birth. Data collected by questionnaires, as described above, were used to create the variables about daily smoking (yes/no) and frequency of eating breakfast during a normal school week (0–1 days/2–3 days/4–5 days per week). Physical activity levels were categorised as low, medium and high active. Low active was defined as moderate + hard physical activity
