 |
 |
 |
 |
||
 |
||
| Home | Command Papers | House Papers | Departmental Papers | Search | Site Map | Contact Us | Links | ||
 |
|
|
|
The
Health of Children and Young People
 9 Anthropometric measurements, overweight, and obesity
Emmanuel Stamatakis
9.1 Introduction Overweight and obesity are terms that refer to an excess of body fat and they usually relate to an increased weight-for-height. The two terms, however, denote different degrees of excess adiposity, and overweight can be thought of as a stage where an individual is at risk of developing obesity. 1 The adverse health consequences associated with obesity are related to an increased adiposity rather than an increased weight per se 2 and it is therefore important that any indicator of obesity reflects this increased adiposity. Body mass index (BMI) takes into account weight and height: it is calculated as weight (kg) divided by squared height (m2) and it is the key overweight and obesity measure in this chapter. BMI has been shown to strongly correlate with adiposity in adults3,4 and children.2,5,6,7 The selection of BMI is further supported by recommendations made by the International Obesity Task Force, which concluded that BMI is a reasonable measure of body adiposity in children.8 Despite the relatively wide acceptance of the use of BMI as an adiposity indicator, the establishment of a specific young obesity and overweight classification system has proved to be difficult. Constant changes in body composition during growth mean that the relationship between weight-for-height and adiposity during childhood and adolescence is age-dependent, and this relationship is further perplexed by race and gender.9 Considering these issues, in addition to the lack of a strong link between adiposity and specific disease end-points (as opposed to adulthood where a clearer relationship between disease end-points and BMI status exists10,11), it is not surprising that defining young overweight and obesity is highly problematic. Several methods have been employed to define early overweight and obesity, including ideal body fatness as measured by skinfold thickness,12,13 national BMI percentile charts,14,15,16 weight-for-height indices,17 and BMI percentile cut-off points.1,18 BMI percentile cut-off points that have been proposed to define obesity are the 95th1 or the 97th percentile,18 while the 85th1 or the 90th18 percentile have been proposed to define overweight. However, the rationales for all above definitions are weak19 and as a result none of them has enjoyed general acceptance. The International Obesity Task Force20 has recently proposed a solution that is based on BMI reference data from six different countries around the world (over 190,000 subjects in total aged 0-25 from UK, Brazil, Hong Kong, The Netherlands, Singapore, and the United States). In summary, the BMI percentile curves that pass through the values of 25 and 30 kg/m2 (standard adult cut-off points for overweight and obesity, respectively) at age 18 were smoothed21 for each national dataset and then averaged. The averaged curves were then used to provide age and sex-specific BMI cut-off points for children and adolescents aged 2-18. By averaging the distribution curves from each reference country, the International cut-offs for children purport to be representative of the countries but independent of the overweight or obesity level in each country.20 One of the benefits of using these International standards is the possibility of making international comparisons. The cut-off curves22 for obesity and overweight for ages 2 to 16 using this classification (subsequently referred to as the International classification) are presented in Figure 9A. This attempt to link childhood and adult obesity/overweight standards was justified on the grounds of the clear associations that have been found between the adult BMI cut-off values of 25 and 30 kg/m2 and health risk.23 Recent evidence24 lends support to the use of the International classification by suggesting links between children's health risk factors (including blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, and glucose) and overweight or obesity (as defined by the International classification). However, the International classification is not without problems: International reference data differ from those for the UK population, and this is reflected in the sex-specific overweight and obesity estimates produced by the International classification. Recent re-analysis25 of children's BMI data using similar methodology to the International classification but UK-only reference data15 showed that the International BMI cut-offs exaggerate the differences in overweight and obesity prevalence between boys and girls by under-estimating prevalence in boys. Other possible limitations of the International classification include concerns about its sensitivity (ability to identify all obese children as obese), the limited sample size of the reference population, and the lack of BMI cut-off points for underweight.26 Despite these and other26,27 criticisms of the epidemiological and clinical value of this approach, the use of the International classification is gaining acceptance.19,28,29,30,31,32,33 However, the issue of childhood obesity definition is far from resolved and there is an urgent need for further work. The main child overweight and obesity prevalence estimates in this chapter have been produced using the International definition. However, in the light of the lack of consensus on its use, key results (overweight and obesity prevalence and prevalence trends) have also been produced using 85th the (overweight cut-off)/ 95th (obesity cut-off) BMI percentiles of the UK reference curves (referred to as the National BMI percentiles classification).15, 16 The National BMI percentiles classification has been used in the past to describe childhood overweight and obesity prevalence trends in UK34,35,36 and was used to produce obesity trends estimates in the Chief Medical Officer's 2002 Annual Report.37 However, the National BMI percentiles classification were not selected as the primary measure in this report as they are based on the arbitrary assumption that the prevalence of overweight and obesity at the point when the reference data15 were compiled was 15% and 5%, respectively. Furthermore, there seems to be no indication that these cut-off points relate directly or indirectly to any physiological outcomes or health or disease risks.38 It is worth noting that the UK component of the International classification used the same sample as that used to construct the UK reference BMI data.15,16,20 Figure 9A The importance
of studying early obesity and overweight is highlighted by the increasing
body of evidence that links obesity with numerous long-term and immediate
health risks. Childhood and adolescent obesity can persist into adulthood,
where the direct health risks of obesity are severe and well-established.10,11
It has been estimated that up to 50% of obese adolescents remain obese
in adulthood.13 Other studies have linked childhood and adolescent
fatness directly to middle-age mortality and morbidity.39,40,41,42,43,44,45
In addition to the increased risk for health problems in later life, children
face the immediate health consequences of obesity, including increased
risks for abnormal lipids profile and elevated blood pressure.46
A significant association between systolic blood pressure and BMI is reported
in the blood pressure chapter of this report47 and other studies.48
Associations between childhood obesity and increased asthma prevalence49
or the incidence of type 2 diabetes mellitus50 have also been
reported. In terms of life quality, severely obese children and adolescents
have been reported to score lower than cancer patients of similar age
in health-related quality of life scales.51 Some studies have reported negative psychological consequences associated with obesity or increases in body mass among children. These consequences include psychological morbidity, more negative physical self-perceptions, unfavourable changes in physical activity and attitudes towards it, and lower scores on general self-esteem, global self-worth, or self-concept.52,53,54 Other studies, on the other hand, have not found these consequences.55,56,57 For example, one study indicates that no psychiatric symptoms were present among obese adolescents.55 Various studies of preadolescents and adolescents have shown that a sizeable proportion are dissatisfied with their body size and shape and want to be thinner. This concern increases with body mass and is more prevalent in females than in males.58,59,60 A study of college students in the US found that females tended to perceive themselves as overweight when they were not, and failed to see themselves as underweight when they were. Many of those who did not see themselves as even slightly overweight wanted to lose weight. Although males reported some dissatisfaction with their bodies, they tended to want to gain rather than lose weight.61 The anthropometric measures and overweight and obesity estimates for children aged under 16 and for young adults (aged 16-24) in this report are based on the combined 2001-2002 Health Survey for England (HSE) datasets. The length of infants whose age at last birthday was less than 2 was measured. For those aged 2 and over, height and weight were measured and BMI was calculated. Infant length, height, weight, and BMI are presented by individual years of age. Overweight and obesity prevalence trends for ages 2-24 were calculated for HSE years 1995 to 2002. For this purpose HSE 2001 and HSE 2002 were considered separately, while years 1999 and 2000 were combined due to their small sample sizes. As already stated above, the prevalence trends 1995-2001 were produced using both the International and the National BMI percentiles classification, based on the UK 1990 reference data.15,16 Previous studies of the relationship between socio-economic classifications and the prevalence of overweight or obesity in childhood and among adolescents have produced conflicting results. While negligible associations were found in some studies,62,63 others found significant socio-economic status variation in the prevalence of overweight or obesity.64,65,66,67,68 The relationships of BMI and overweight and obesity prevalence with a number of selected socio-economic and demographic variables have been examined in this chapter. Such variables included socio-economic classification (NS-SEC), area deprivation levels, equivalised household income, area type, and Government Office Region. Another set of psychological and behavioural variables that may be associated with BMI were also examined, namely perceived body weight, attempted weight change, and psychological well-being. Logistic regression models were developed where the relationship of overweight (including obesity) and obesity only, with a number of other variables was examined. These included some of the socio-demographic variables mentioned above (equivalised household income, area deprivation levels, socio-economic classification) in addition to birth weight, physical activity, fruit consumption, household size, and parental BMI status. All analyses in this chapter are cross-sectional, and cannot therefore link variables in a 'cause' and 'effect' manner. Instead, the results presented below can only highlight relationships or lack of relationships between variables. The terms 'overweight' and 'obese' in this chapter are mutually exclusive, 'overweight' excluding those who are obese. Where a statement is made about the aggregate of these two categories, terms such as 'overweight including obesity' or 'overweight including the obese' are used. 9.2 Response to anthropometric measures The large majority of informants who had a nurse visit provided valid weight and height measurements: it was possible to compute BMI for 88% of boys and girls aged 2-15, for 93% of young men, and for 91% of young women aged 16-24 who had a nurse visit. In total 4740 boys, 4672 girls, 2308 young men and 2663 young women had a valid BMI measurement. Additionally, 83% of male and 84% of female infants aged 0-1 had a valid length measurement (data not shown). In total 377 male and 382 female infants aged 0-1 had a valid length measurement. Response to height and weight measures was markedly lower for those aged 2-3 than for those aged 4 and over. Overall, response to all anthropometric measures was slightly higher for young adults than for children. 4 young men and 2 young women were not classified as having a valid weight measurement as the scales used could not register weights over 130 kg. However, they were included in the BMI analyses as they gave their estimated weights. Table 9.1 9.3 Anthropometric measures by age and sex 9.3.1 Height From age 2 to age 11, height was almost identical for boys and girls. Boys were very slightly taller in the earlier years (by about 1%). This difference decreased with age until ages 11 and 12, when girls' mean height slightly exceeded that of boys. After age 13 the rate of height increase slowed down for girls, levelling off at around age 15. Boys in contrast continued to grow for a further three years before their height similarly levelled off, at around age 18, by which time they were 8% taller than girls (the mean height of boys aged 17 was 176.6 cm, of girls 162.9 cm, a difference of 13.7 cm). From age 2 to age 13 for girls and age 15 for boys, height was an approximately linear function of age, as can be seen in Figure 9B. At any given individual age year, height was approximately normally distributed. Tables 9.2, 9.3, Figures 9B, 9C 9.3.2 Weight Whereas mean height increased linearly with age up to age 13 (15 for boys), the increase in mean weight by age, as shown in Figure 9D, has an S shaped curve, the rate of increase initially approximating to linear (to about age 8), and then accelerating until around age 14 for girls (18 for boys), after which it slowed down. At age 24, mean weight was 78.8 kg for males, 66.2 for females. Taken in isolation, the data suggest that by age 24 mean weight has reached its peak, but earlier data covering the whole adult population shows that mean weight continues to increase (although slowly) beyond this age.69 At age 3, boys were on average about 4% heavier than girls, but this percentage gradually declined until age 12, when the situation was reversed, with boys on average being 6% (2.9 kg) lighter than girls. Thereafter, boys' continuing growth led to an increasing gap between their mean weight and that of girls. By age 24 they were about 20% (about 13 kg) heavier, compared with 8% taller. Weight was positively skewed, slightly more for females than males. At any given individual age year, means were typically 2% to 6% higher than medians. The distribution of weight at any age had a much larger variance than the distribution of height at that age. In both sexes, but particularly in females, the variance tended to increase with age. Table 9.4, Figures 9D, 9E
9.3.3 BMI BMI is the ratio of weight (in kg) to the square of height (in metres). It is shown by year of age in Figure 9F, separately for each sex. In early years, weight increases faster than height, but not quite so fast as height squared, with the well-known consequence that BMI decreases from age 2 to about age 5 or 6 before rising again (since the rate of weight gain subsequently increases more rapidly than the increase in height squared). This is the 'adiposity rebound.70,71 The adiposity rebound was more marked among males than females. Weight gain continued after height (and height squared) levelled off in the middle to late teens, and BMI thus continued to rise. In the present data set it appeared to peak at age 23 for boys and 21 for girls, just before the upper age limit of the sample was reached, but an examination of adult BMI in other HSE datasets shows BMI continuing to increase until middle age.72 The distribution of BMI in males had a slight positive skew. In females it also had a slight positive skew that increased with age. The variability of BMI, as indicated by the standard deviations of the mean, increased with age from about age 7 for boys and from about age 5 for girls. Female informants had marginally lower BMI scores than males up to the age of 6, then higher up to the age of 17. Differences were generally small, with the exception of ages 12 to 16 when differences ranged from about 1.0 to 1.4 kg/m2. Table 9.5, Figures 9F, 9G
9.4 BMI, overweight and obesity among children and young adults 9.4.1 Variations in BMI, and in overweight and obesity prevalence, by demographic and socio-economic indices Overweight and obesity prevalence for children 2-15 was estimated using both the age and sex-specific International classification standards20 and the National BMI percentiles classification that have been discussed in the introductory sections of this chapter. The calculations for the International classification estimates were produced using the published table with the age (at 0.5 years intervals) and sex-specific BMI cut-off values22 for overweight and obesity. The National BMI percentiles classification estimates were produced by calculating the percentage of boys and girls who were over the 85th (overweight) or over the 95th (obese) BMI percentiles of the 1990 reference population. For all overweight and obesity prevalence estimates the exact age was used (extracted from the interview date minus the date of birth). Presentation of the results is based, however, on the age at last birthday which is the HSE standard. For young adults aged 16-24, prevalence estimates were made using the standard adult BMI definition of overweight (over 25 to 30 kg/m2) and obesity (over 30 kg/m2). The decision not to use the International classification standards or the National BMI percentiles classification for those aged 16 to 18 was taken on the grounds of the need for continuity with previous Health Survey reports, where the standard adult definition was used for informants 16 and over. The standard adult definition provided moderately lower estimates than the International definition for ages 16 and 17 but identical figures for age 18, as the International BMI cut-off points for overweight and obesity are the same as the standard adult. The table below shows the different overweight and obesity estimates derived by the International and the standard adult definitions.
By age and sex (International classification) The top section of Table 9.6 presents the obesity and overweight prevalence estimates for children derived using the International classification, by age and sex. Obesity prevalence in children 2-15 was 5.5% for boys and 7.2% for girls (p<0.01). In total, over a fifth of boys (21.8%) and over a quarter of girls (27.5%) were either overweight or obese. The sex difference in combined overweight plus obesity prevalence was statistically significant (p<0.0001). The pattern of overweight and obesity prevalence in individual age years was irregular for children 2-15, but rates were consistently higher for girls than for boys. Figure 9H shows moving averages (of three age years) for obesity and overweight (including obesity) prevalence. Obesity and overweight (including obesity) prevalence percentages were higher in females than in males throughout childhood. Differences in obesity prevalence between the sexes were pronounced for ages 4-9 (p<0.0001). However, the higher overweight and obesity prevalence figures for girls than for boys across childhood should be viewed with caution since it has been suggested that the International classification may exaggerate the sex differences by under-estimating prevalence for boys.25 Obesity prevalence rates were higher in females than in males in young adulthood, and obesity rates among young women 19-21 were higher than among young men of this age by 4 to 6 percentage points. Obesity prevalence in young adults 16-24 was 9.2% for young men and 11.5% for young women and although this sex difference is small, it was statistically significant (p<0.01). Overall, about a third of young men (32.2% in total) and young women (32.8% in total) were classified as overweight or obese. Table 9.6, Table 9.7, Figure 9H
By age and sex (National BMI percentiles classification) The bottom section of Table 9.6 presents the overweight and obesity prevalence estimates for children aged 2-15 derived using the National BMI percentiles classification, by age and sex. In comparison with the International classification, obesity estimates derived by the National BMI percentiles classification are much higher (16.0% vs 5.5% for boys and 15.9% vs 7.2% for girls), but overweight estimates are lower (14.3% vs 16.3% for boys and 14.8% vs 20.3% for girls). Differences between the two definitions are small for girls when the combined overweight including obesity category is considered (30.7% vs 27.5%) but remains more marked for boys (30.3% vs 21.8%). In contrast to the International classification, no marked sex differences were observed in overweight or obesity prevalence using the National BMI percentiles classification. Table 9.6 By equivalised household income Table 9.8 presents age-standardised mean BMI, and overweight and obesity prevalence, by equivalised income quintiles. Age standardisation was performed to take into account the age differences between the different categories of the equivalised income quintiles. The standardization reference was the 2001 population figures (based on 2001 Census). The same standardisation method was applied to all other socio-economic and demographic classifications presented in this section (Tables 9.8 to 9.12). Earlier HSE reports have found an inverse association between adult obesity and equivalised household income. For example, in the 1998 survey, the prevalence of obesity ranged from 14.8% (men) and 15.2% (women) in the highest income quintile to 20.1% (men) and 26.0% (women) in the lowest income quintile.72 Among children and young adults in HSE 2002, an inverse relationship was found between obesity prevalence and equivalised income quintile (i.e. decreasing prevalence rates while moving from lower to higher income quintiles). This relationship was stronger for females than males (it was weaker among boys than girls, and was not present in young men). The proportions obese in the highest and lowest income quintiles respectively were 4.2% and 6.0% for boys, 3.8% and 8.3% for girls (p<0.001), 12.7% and 10.3% for young men, and 8.4% and 13.1% for young women. The proportion who were either obese or overweight also varied among girls from 24.5% in the highest income quintile to 28.6% in the lowest, but no systematic variation was found among boys, or among young adults of either sex. Table 9.8 By socio-economic classification The socio-economic group of the household reference person of each participating household was defined using the new National Statistics Socio-economic Classification (NS-SEC) which was introduced for the first time in the UK Census 2001. Informants were assigned to 14 functional and 3 residual NS-SEC categories on the basis of the current or former occupation of the household reference person. These were then collapsed to the five-category version defined as follows: i) managerial and professional, ii) intermediate (e.g. clerical, administrative, sales), iii) small employers and self-employed workers, iv) lower supervisory and technical, and v) semi-routine and routine occupations.73 Obesity prevalence was lower in households with a reference person in a managerial or professional occupation than in those where the reference person was in a routine or semi-routine occupation. For boys the difference was 3.8% to 5.6%, for girls 5.0% to 9.0% , for young men 8.7% to 11.9%, for young women 7.6% to 14.8%. As in the case of equivalised household income, the association was stronger for females than males. These differences between the first and last NS-SEC groups do not necessarily imply a smooth gradient from one to the other. For example, among young adults, where the reference person was in an intermediate occupation, obesity prevalence was higher than where he or she was in a routine or semi-routine occupation. Table 9.9 The relationship between overweight and obesity and NS-SEC of the household reference person did not change when estimates for children 2-15 were calculated using the National BMI percentiles classification. As shown in the table below, overweight (including obesity) prevalence was lower among boys from households where the reference person was in a managerial or professional occupation compared with all other NS-SEC categories.
By area deprivation The Index of Multiple Deprivation 2000 (IMD) is based on levels of deprivation in local areas, in six domains (income; employment; health deprivation and disability; education, skills and training; housing; geographical access to services).74 The index is presented in quintiles ranging from the first (the least deprived) to the fifth (the most deprived). Differences in obesity prevalence were observed between the first and last quintiles. Among boys, obesity prevalence was 4.5% in the least deprived quintile, 6.4% in the most deprived (p=0.05). Corresponding figures for other groups were girls 4.5% and 8.8% (p<0.0001), young men 8.9% and 12.5%, and young women 7.6% to 14.1% (p<0.001). As in the case of equivalised household income and NS-SEC, the differences were greater among females than males. When the two least deprived quintiles were aggregated versus the aggregate of the three most deprived quintiles, mean BMI and overweight and obesity prevalence were found to be generally higher in the most deprived than in the least deprived quintiles. These differences were statistically significant in most age/sex groups except for boys aged 2-15, as the table below shows. Differences in percentages are shown in percentage points (pp): for example, the difference in obesity prevalence in the case of boys is 4.9% to 5.8%, which is 0.9 percentage points. Table 9.10
By area type Area types were classified by the interviewer as one of the following: inner city, other dense urban/town centre, suburban residential, rural residential/village centre, rural agricultural with isolated dwellings or small hamlets. No association was discernible between area type and mean BMI or overweight prevalence. The prevalence of obesity in boys and girls aged 2-15 was higher in inner cities than in other types of areas. Among young adults, however, obesity prevalence was relatively low in inner city areas. Table 9.11 By Government Office Region There are currently nine Government Office Regions:75 North East, North West, Yorkshire and the Humber, East Midlands, West Midlands, East England, London, South East, and South West. There were isolated instances of raised obesity or overweight prevalence in some regions, but overall no clear regional patterns in mean BMI or obesity and overweight prevalence were identified either among children or among young adults. Table 9.12 9.4.2 Factors associated with overweight in children aged 2-15 To link BMI status with possible risk factors, two multivariate logistic regression models were developed. In one, the dependent variable was being in either the overweight or the obese category of the International BMI classification standards for children.20 In the other, the dependent variable was being obese, again according to the International classification. Separate logistic regression models were developed for boys and girls in households where both parents were present and both parents had valid BMI measurements. The initial models for both sexes resulted in all variables being left out of both models as non-significant, with the sole exception of parental BMI status. This involved four categories: both parents obese, father but not mother obese, mother but not father obese, neither obese (the reference category). The results of the final models, which included only this single significant independent variable, are presented in Tables 9.13 and 9.14. Similar logistic regression models were not developed for young adults aged 16-24 as no data were available on birth weight or parental BMI for this group. Dependent variable: overweight or obese The risk for being overweight or obese as a child increased with the number of obese parents. Boys were over six and a half times as likely to be overweight or obese if both parents were obese (BMI>30 kg/m2). They were also two times more likely to be overweight or obese if they had an obese father and three times more likely if they had an obese mother. Similarly, girls were about seven times as likely to be overweight or obese if both parents were obese. Girls were also about two times more likely to be overweight or obese if their father or mother was obese. Strong links between children's weight and parental weight status have been reported in HSE 199776 and in other cross-sectional77,78 and longitudinal79 studies. Table 9.13 Dependent variable: obese Re-running the model using membership of the obesity category as the dependent variable again found the number of obese parents to be the only significant independent variable. However, the odds of being an obese child were considerably higher than the odds of being an overweight child (as defined above) if the mother, or both parents were obese, as shown in the summary table below: Table 9.14
9.5 Trends over time in height, weight, BMI, overweight and obesity among children and young adults 9.5.1 Introduction Trends in anthropometric measures and in obesity/overweight were assessed on a year-by-year basis using the individual HSE datasets 1995-2002. HSE 1999 and HSE 2000 datasets were combined due to the small sample size of each individual year. To account for the fact that a maximum of two children were selected from each household, data were weighted using the weighting variables of each individual year. Data were not weighted for those aged 16-24. Trend data are presented in age groups that indicate different adiposity growth periods of the childhood: ages 2-5 represent the period prior to adiposity rebound, years 6-10 correspond to the period from the adiposity rebound to approximately the onset of puberty, and years 11-15 represent the years of puberty. Such an age grouping can only be an approximation since the onset of puberty is subject to inter-individual variation and generally occurs earlier for girls than for boys. However, an accurate assessment of individual pubertal status was not possible as pubertal stage data were not available. The International BMI classification standards20 described earlier in this chapter were used for the assessment of the overweight and obesity prevalence trends in this section. The robustness of the trends elicited using these standards was tested against the trend estimates produced by the National BMI percentiles classification. For all overweight and obesity prevalence trend analyses the exact age was used (extracted from the interview date minus the date of birth). Presentation of the results is based, however, on the age at last birthday which is the HSE standard. The distribution of children's ages in the population, and in the HSE samples, has increased over time, as illustrated in the table below. This could be a confounding factor when comparing growth-sensitive data, such as height, weight, and BMI across years. For this reason trend data were age-standardised using the individual age year proportions of boys, girls, young men, and young women from the 2001 Census. All trend data presented in wide age bands (i.e. 2-15 and 16-24) in this section and in the relevant tables and figures have been age-standardised using this scheme. No standardisation was performed for data presented in narrow age bands (e.g. Tables 9.15-9.17).
9.5.2 Height and weight trends 1995-2002 No discernible height trend was evident among male and female informants of any age group. Some increases between 1995 and 2002 were found among young boys aged 2-5 (increase by 1.0 cm) and 6-10 (increase by 0.5 cm), and among girls 6-10 (increase by 0.4 cm) and 11-15 (0.2 cm increase). In contrast, the mean height of women 16-19 decreased by about one cm. On the whole, the height of boys aged 2-15 increased marginally, while girls' height decreased marginally between 1995 and 2002 (see table: data shown only for children).
In contrast to height, where no strong trend emerged, an upward trend was evident for weight across all age and sex groups. The table below presents the age-standardised totals for children's and young adults' weights between 1995 and 2002. Increases in weight between 1995 and 2002 were greater among young adults (increase by 1.6 kg and 2.3 kg for young men and young women, respectively) than among children (increase by 1.1 kg and 0.8 kg, for boys and girls, respectively). For both male and female informants, the age group that presented the greatest increases was 20-24 years (increase of 1.8 kg and 3.2 kg for males and females, respectively). Table 9.15, Table 9.16
9.5.3 BMI, and overweight and obesity prevalence trends 1995-2002 BMI trends Gradual increases in mean BMI between 1995 and 2002 were evident for most age groups of children and young adults. Increases were more marked for children aged 6-10 and 11-15, and for young adults aged 20-24. The table below shows that, despite fluctuations within age groups, there was a steady upward trend in age-standardised mean BMI for boys, girls, young men and young women. Table 9.17
Overweight and obesity prevalence trends (International classification) Table 9.18 presents year-by-year overweight and obesity prevalence trends, for children (using the International classification cut-offs) and for young adults. A consistent upward trend was evident across all age groups in children 2-15. In line with the BMI trends, the obesity and overweight prevalence upward trend was clearer for age groups 6-10 and 11-15. In total, between 1995 and 2002 obesity prevalence almost doubled among boys aged 2-15 from 2.9% to 5.7%, an increase of 2.8 percentage points (pp), while the increase among girls of the same age group was 2.9 pp (from 4.9% to 7.8%). The increase in overweight (including obesity) prevalence over the period of seven years was 5.6 pp for boys and 6.0 pp for girls. Prevalence trends for young adults aged 16-24 between 1995 and 2002 were characterised by a marked increase in the prevalence of obesity. Obesity prevalence among young men aged 16-24 increased by 3.6 pp between 1995 to 2002 (from 5.7% to 9.3%). Similar upward trends were observed among young women aged 16 and 24 where obesity prevalence rose by 3.9 pp and overweight (including obesity) prevalence increased by 6.5 pp. Table 9.18, Figure 9I Overweight and obesity prevalence trends (National BMI percentiles classification) Table 9.19 presents the results of the prevalence trend analysis for children aged 2-15 using the National BMI percentiles classification standards, by age group and sex. Although the figures produced by the two definitions differ considerably (as described in Section 9.4.1), the actual trend is not affected by the definition used. The National BMI percentiles classification trends are in agreement with the trends produced using the International BMI cut-off points in that marked increases have occurred in the prevalence of obesity, while the prevalence of overweight has increased only moderately. Table 9.19, Figure 9J Two other studies28, 33 that used the International
classification20 reported increasing trends in the prevalence
of overweight and obesity. A very recent report33 that used data
from the National Study for Health and Growth (NSHG) and from HSE documented
markedly sharp increases in overweight and obesity prevalence between 1994
and 1998 for English boys and girls 5-17. Another study28 that
used data from the NSHG assessed overweight and prevalence trends over 30
years among English and Scottish children from age 4 to age 11.
Although little difference
was found between 1974 and 1984, the differences between 1984 and 1994
were marked: overweight increased on average, for English boys and girls
respectively, by 3.6 pp and 4.7 pp, while obesity increased by 1.2 pp
and 1.4 pp. NSHG has also provided data on skinfold thickness trends,80 and it has been shown that between 1972 and 1994 the skinfold thickness of English and Scottish children aged 5-11 increased on average by 11.1 pp and 10.3 pp for boys and girls respectively. Data collected routinely by health visitors in the North West of England have shown also that from 1989 to 1998 there was an 8.9 pp increase in overweight (defined as >=85th BMI percentile of the 1990 UK reference data15) and a 3.8 pp increase in obesity (defined as >=95th BMI percentile of the 1990 UK reference data15) within a large sample of English infants and pre-school children.35 Other smaller scale studies have confirmed this upward trend in obesity and overweight prevalence.34, 36 The above evidence highlights that the increasing trend in children's obesity and overweight prevalence is independent of the method used to define childhood obesity. Despite the difference methodologies used among studies, evidence suggests that obesity is increasing in children and that the rate of this increase is high. The upward trend in childhood overweight and obesity prevalence is not confined to England only. Numerous recent studies around the world have used the International classification standards to assess overweight and obesity prevalence trends in childhood, and all identified studies reported increases in the prevalence of overweight and obesity over time.81,82,83,84,85,86,87 For example, the prevalence of overweight (including obesity) among Australian boys and girls aged 7-15 doubled, and that of obesity alone tripled, between 1985 and 1997.81 Another study reported considerable increases in the prevalence of overweight (including obesity) in Brazil (1975-1997), China (1991-1997), and the United States (19711974 and 1988-1994) among children and adolescents 6-18.82 Similar findings from studies that used the International classification standards have been reported for those aged 5-12 in France (1992-2000),83 8-14 in East Germany (1992-1999),84 5-6 in West Germany (1982-1997),85 12-18 in Finland (1977-1999)86 and 6-13 in Sweden (1986-2001).87 An analysis of trends in HSE by socio-economic group is limited by the fact that NS-SEC was not used before 2001, Registrar-General's social class having been used in earlier years. Social class was coded on the 2001 and 2002 datasets, and can thus be used for trend analysis. The six social class categories were collapsed into two:
non-manual (I, II, and IIINM) and manual (IIIM, IV, and IV). The following
table below presents the obesity prevalence trends for various age groups
for aggregated manual and non-manual classes. The International classification
was used to estimate obesity in children 2-15 in this analysis. Although
there was an upward trend for both the manual and non-manual classes in
the prevalence of obesity over time, increases were more pronounced among
manual classes across all age and gender groups. Considering males and
females together, obesity prevalence increased from 5.4% in 1995 to 9.4%
in 2002 (4.0 pp) for the aggregated manual social class categories and
from 4.0% in 1995 to 6.4% in 2002 (2.4% pp) for the aggregated non-manual
social classes.
9.6 BMI, and overweight and obesity, in relation to psychological indices 9.6.1 Perceived weight, attempted weight change and obesity Perceived weight Informants 8-15 were asked 'Given your age and
height, would you say that you are about the right weight, too heavy,
too light, or not sure?' Responses are summarised in the table below.
For both sexes, self awareness of body weight increased markedly with age. For males and females respectively, the percentage who were not sure whether their body weight was about right decreased from 22% and 27% at age 8-15 to 11% and 12% at age 16-24. In both sexes and both age groups the majority felt that their weight was about right. However, over a third (37%) of young women aged 16-24 tended to believe that they were too heavy, a considerably higher proportion than of young men of the same age (19%). About one in six (16%) young men aged 16-24 believed that they were too light, compared with about one in seventeen young women (6%). Perceived weight, by BMI quintile Perceptions of weight were related to BMI. The proportion who believed that they were too heavy increased with increasing BMI. Increases with BMI were steeper among young adults than among children. The proportion believing that they were too light decreased with BMI quintile. Perceived weight, by overweight and obesity status Substantial proportions of overweight boys (42%) and girls (37%) aged 8-15 believed that their weight was about right. A similar proportion of overweight young men aged 16-24 held this view (47%), but among overweight young women the proportion was smaller (19%). There was a gradual increase with age in the proportion of females who regarded themselves as too heavy, despite their being neither overweight nor obese: in total, 20% of young women aged 16-24 who were neither overweight nor obese perceived themselves as too heavy. Awareness of being too heavy increased progressively from 42% for obese boys aged 8-9 to 81% of obese young men aged 20-24; and from 31% of obese girls aged 8-9 to 97% of obese young women aged 20-24. Tables 9.20, 9.21 Attempts to change weight A question was asked also about attempts made to change weight: 'At the present time, are you trying to lose weight, trying to gain weight, or are you not trying to change your weight?' As shown in the table below, the proportion trying to lose weight increased with age, particularly among female informants: about a third (31%) of girls aged 8-15 were trying to lose weight, increasing to more than half (54%) for young women aged 16-24. In contrast, there was a sizeable increase with age in the percentage of men who were trying to gain weight (from 9% for those aged 8-15 to 21% for those aged 16-24). The increase in self-awareness of weight with increasing age, reported in the previous section, was reflected in a decrease in the proportion not trying to change their weight: for males and females respectively, from 73% and 65% at age 8-15 to 57% and 42% at age 16-24.
Attempts to change weight were associated with BMI. The percentage who wanted to lose weight increased with BMI quintile, while the percentage who wanted to gain weight decreased with BMI quintile. It is noteworthy that one in 10 (10%) of young women aged 16-24 in the bottom BMI quintile, and more than one in three (36%) of young women in the second BMI quintile, were trying to lose weight. As shown in Table 9.22, the percentages of girls aged 8-15 and young women aged 16-24 who were trying to lose weight despite not being overweight or obese were 19% and 41% respectively; the corresponding proportions for males were 10% and 9%. Substantial proportions of overweight boys (57%), young men (42%), and girls (46%) did not express a desire to change weight. Furthermore, one in three (33%) obese boys and more than one in three (39%) obese young men were not trying to change weight. The corresponding percentages were lower for obese girls (23%) and obese young women (14%). Interestingly, the proportion who were trying to change weight (lose or gain) was higher than the proportion dissatisfied with their body weight, as reported in the previous section. As shown in the table below, this difference was more marked among young women aged 16-24. This may be due to, for example, perceptions of peer group pressures that did not entirely match the informant's self-perception. Tables 9.22, 9.23
9.6.2 Psychological well-being and obesity Psychological well-being was assessed in HSE 2002 by the General Health Questionnaire (GHQ12). GHQ12 was originally designed for use in general practice settings as a screening tool and has been used in HSE every year except 1996. GHQ12 is designed to be self-administered and was included in the self completion booklet for informants aged 13-24. The questionnaire comprises 12 questions, asking informants about their general levels of happiness, experience of depressive and anxiety symptoms, and sleep disturbance over the last four weeks. Interpretation of the answers is based on a four point scale response scale scored using a bimodal method (symptom present: 'not at all'=0, 'same as usual'=0, 'more than usual'=1, 'much more than usual'=1). A GHQ12 score of 4 or more indicates the possible presence of psychiatric morbidity. There was no clear association between BMI and GHQ12 score. Tables 9.24, 9.25
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
