点击显示 收起
1 From the Department of Paediatrics and Child Health, The University of Sydney at The Children’s Hospital at Westmead, New South Wales, Australia (MLB); the Epidemiology Unit, South Western Sydney Area Health Service, Liverpool, New South Wales, Australia (TC); the Centre for Community Child Health, Royal Children’s Hospital, Victoria, Australia (MW and KH); and the School of Physical Education, Exercise and Sport Studies, University of South Australia, Adelaide, Australia (KN, JD, and IR).
2 Supported in part by grants from the New South Wales (NSW) Department of Education and Training, the NSW Department of Health, the Public Health Division of the Victorian Department of Human Services, and the National Professional Development Program. 3 Address reprint requests to ML Booth, Department of Paediatrics and Child Health, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead New South Wales 2145, Australia. E-mail: michaeb4{at}chw.edu.au.
ABSTRACT
Background: Effective public policy requires information on the prevalence of overweight and obesity.
Objective: We determined changes in the population prevalence of overweight and obesity among young Australians (aged 7–15 y) from 1969 to 1985 to 1997.
Design: Data from 5 independent population surveys were analyzed: the Australian Youth Fitness Survey, 1969; the Australian Health and Fitness Survey, 1985; the South Australian Schools Fitness and Physical Activity Survey, 1997; the New South Wales Schools Fitness and Physical Activity Survey, 1997; and the Health of Young Victorians Study, 1997. Measured body mass index was used as the index of adiposity, and recently published body mass index cutoff values were used to categorize each subject as nonoverweight, overweight, obese, or either overweight or obese.
Results: For 1985–1997, the population prevalence of overweight increased by 60–70%, obesity increased 2–4-fold, and the combined overweight and obesity categories doubled. The findings were consistent across data sets and between the sexes. For 1969–1985, there was no change in the prevalence of overweight or obesity among girls, but among boys the prevalence of overweight increased by 35%, the prevalence of obesity trebled, and the prevalence of overweight and obesity combined increased by 60%.
Conclusions: The data show that in 1985–1997, the prevalence of overweight and obesity combined doubled and that of obesity trebled among young Australians, but the increase over the previous 16 y was far smaller. These results should increase our sense of urgency in identifying and implementing effective responses to this major threat to public health.
Key Words: Children • adolescents • overweight • obesity • body mass index • BMI • epidemiology
INTRODUCTION
Overweight and obesity place children and adolescents at increased risk of significant health problems, both during their early life and during adult life (1–3). A clear understanding of the epidemiology of overweight among children and adolescents makes an important contribution to the development of an effective response. Reports on the prevalence and the population distribution of overweight and obesity among Australian children and adolescents were published in the past several years (4, 5), and the present paper adds to the epidemiologic picture by reporting changes in the prevalence of overweight and obesity among young Australians over the period 1969–1997.
Freedman et al (6), drawing on data from the Bogalusa Heart Study, reported that the prevalence of overweight and obesity [defined as > 85th centile of body mass index (BMI)] doubled between 1973 and 1993, that increases in the prevalence of overweight accelerated through this period, and that the largest increases occurred at the higher end of the distributions of body mass index. A study of children in regional Spain also found that, based on the second National Health and Nutrition Examination Survey values for the 95th centile, the prevalence of overweight approximately doubled among boys and girls aged 6–7 y and among boys (but not girls) aged 13–14 y between 1985 and 1995 (7). Chinn and Rona (8) applied Cole et al’s (9) new BMI cutoffs to data on British children collected in 1974, 1984, and 1994. They found no change in the prevalence of overweight or obesity between 1974 and 1984, but the prevalence of overweight increased by 44–67% in 1984–1994 and the prevalence of obesity approximately doubled over the same period. These reports suggest that not only is the prevalence of overweight and obesity increasing but that it has increased more rapidly over the past 10–15 y than in previous decades.
The purpose of the present study was to confirm similar changes in Australian children between 1969 and 1997. Data from 5 surveys conducted in 3 Australian states were used to identify changes over the periods 1969–1985 and 1985–1997. Data collected in South Australia (SA) during the Australian Youth Fitness Survey, 1969; the Australian Health and Fitness Survey, 1985; and the SA Schools Fitness and Physical Activity Survey, 1997, were compared. Comparisons were made between the New South Wales (NSW) Schools Fitness and Physical Activity Survey, 1997, and the NSW sample of the Australian Health and Fitness Survey, and we also compared data from the Health of Young Victorians Survey, 1997, with data collected as part of the Victorian sample of the Australian Health and Fitness Survey, 1985. Absolute criteria for BMI for children and adolescents were applied to each data set, allowing us to determine changes in the prevalence of overweight and obesity among Australian children and adolescents.
SUBJECTS AND METHODS
The 5 data sets used for this study are briefly introduced and the survey methods are summarized in Table 1. All studies included representative samples drawn for state or national surveys, and all were cross-sectional in design.
View this table:
TABLE 1 . Summary of the methods of each study1
The Australian Youth Fitness Survey, 1969, was a national cross-sectional survey (n = 7439) of students selected randomly from government high schools in urban and rural areas (age range: 13–17 y). Multiple measures of strength and aerobic fitness were administered in addition to height and weight. It should be noted that the smaller states were oversampled to provide sufficient numbers for reasonable prevalence estimates, so the sample was not truly nationally representative. However, for the purposes of these analyses, we employed only the data collected in South Australia so that the findings would not be affected.
The Australian Health and Fitness Survey, 1985, was a large-scale national study (n = 8484) that assessed many aspects of health-related fitness, lung function, blood pressure, serum lipids, adiposity, and a broad range of self-reported behaviors among students aged 7–15 y (10). Height, weight, and waist girth were assessed among all children, but skinfold thickness was measured only among students aged 9, 12, and 15 y. Random selection of schools was stratified by state, so data for SA, NSW, and Victoria could be extracted for comparisons with the other data sets, all of which were based on state surveys.
The SA Schools Fitness and Physical Activity Survey, 1997, was a cross-sectional study of the health-related fitness and physical activity of South Australians aged 10–12 y (n = 1904). Measures of a range of fitness attributes, anthropometry (height, weight, girths, and skinfold thicknesses), and self-reported physical activity were collected in 28 randomly selected primary schools throughout urban and rural South Australia (12).
The NSW Schools Fitness and Physical Activity Survey, 1997, surveyed 5518 students in school years 2, 4, 6, 8, and 10 (13, 14). Height, weight, waist and hip girths, a set of measures of health-related fitness, and tests of competency at 6 fundamental movement skills were administered to students in years 4, 6, 8, and 10. Only height and weight were measured among year 2 students.
The primary school subset of the Health of Young Victorians Study, 1997, included height and weight data for 3104 children aged 5–13 y in 24 primary schools across Victoria. Children were measured at school, and parents provided sociodemographic information.
Cole et al’s (9) recently developed age- and sex-specific BMI cutoffs for overweight and obesity among children and adolescents were applied to the data. They provide values of BMI for males and females aged 2–18 y in 6-mo intervals. Ages of the subjects were calculated on the basis on birth dates (self-reported for older subjects or reported by parents or extracted from school records for younger subjects) and the date of participation in the survey, with the exception of the Australian Health and Fitness Survey, 1985. Participants in this survey aged 7 and 8 y reported only their age in integer years, so the value of their BMI cutoff was calculated from the midpoint of their reported age. For example, if a child’s age was 8 y the value of his or her BMI cutoff was taken as the value provided for children aged 8.5 y.
Statistical analysis
Data analyses were carried out with SAS for WINDOWS version 8.1 (SAS Institute, Cary, NC) and SUDAAN version 7.5.2 for personal computers (Research Triangle Institute, Research Triangle Park, NC) software. The number and prevalence of each obesity classification were tabulated separately by age and sex. Multiple logistic regression was used to adjust for exact calendar age and design effects in calculating the odds of overweight and obesity combined by year of survey with reference to the earlier survey.
RESULTS
Of the schools approached to participate in the Australian Health and Fitness Survey, 1985, 90.1% agreed, and, of the students invited to participate, 67.5% agreed to join the study.
South Australia
Response rates for the Australian Youth Fitness Survey, 1969, were not reported, but because nonparticipants were not systematically replaced, a response rate of 84% can be deduced from a target sample of 8900 and a completed data set comprising 7439 children. For the SA Schools Fitness and Physical Activity Survey, 1997, 28 schools (85% of those invited) agreed to participate. The overall student response rate within schools was 72%.
The prevalence of nonoverweight, overweight, obesity, and overweight and obesity combined in 1985 and 1997 is shown (separately for boys and girls aged 10, 11, and 12 y) in Table 2. The table also shows (separately for boys and girls aged 13, 14, and 15 y) the prevalence of nonoverweight, overweight, obesity, and overweight and obesity combined for 1969 and 1985.
View this table:
TABLE 2 . Proportion of boys and girls in each BMI category, including overweight and obese combined (O + O), for the 1969, 1985, and 1997 South Australian data sets
Among both boys and girls, the proportions who were overweight or obese increased significantly between 1985 and 1997 for all ages combined, with increases found within single-year brackets for those aged 10 and 11 y but not 12 y. The odds ratios for all ages combined indicate that the prevalence of overweight and obesity approximately doubled between 1985 and 1997 for both sexes.
In contrast, for both boys and girls there were no statistically significant differences in the prevalence of overweight and obesity combined between 1969 and 1985 for those aged 13, 14, or 15 y. However, there was a significant increase for all boys combined, the odds ratio showing that the prevalence of overweight and obesity combined increased by 60%. There was no increase in the prevalence of overweight and obesity combined among girls over the same period.
The mean height and weight for boys and girls for each survey and the mean waist girth and sum of 3 skinfold thicknesses (biceps, triceps, and subscapular) where they were available are shown in Table 3. Among boys, mean weight increased by 1.1 kg and 2.8 kg and mean height increased 1.2 cm and 1.9 cm for 1969–1985 and 1985–1997, respectively. The mean waist girth increased by 3.8 cm and the mean sum of 3 skinfold thicknesses increased by 5.2 mm for 1985–1997. Among girls, the mean weight did not change over the period 1969–1985, but the mean weight increased by 2.0 kg for 1985–1997. The mean height did not change for girls over the earlier period, but the mean height increased by 1.2 cm over the latter period. The mean waist girth increased by 4.2 cm and the mean sum of 3 skinfold thicknesses increased by 6.6 mm for 1985–1997.
View this table:
TABLE 3 . Mean weight, height, waist girth, and sum of 3 skinfold-thickness changes, 1969–19971
New South Wales
In the 1997 NSW Schools Fitness and Physical Activity Survey, response rates for boys and girls in each primary school year were all > 90%, and for year 8 boys and girls and year 10 boys they were > 80%; for year 10 girls the response rate was 71%. The majority of cases of nonparticipation (> 70%) were due to absenteeism on the day of testing rather than to refusal to participate, although it is recognized that absenteeism may be a passive form of refusal to participate.
The prevalence of nonoverweight, overweight, obesity, and overweight and obesity combined for the NSW component of the Australian Health and Fitness Survey and the NSW Schools Fitness and Physical Activity Survey, 1997, is shown (separately for boys and girls aged 7, 9, 11, 13, and 15 y) in Table 4. Among boys, the proportions who were overweight or obese increased significantly between 1985 and 1997 for every age group. The odds ratios were all statistically significant and indicate that the prevalence of overweight and obesity combined doubled between 1985 and 1997. Among girls, the prevalence of overweight and obesity increased significantly for those aged 11, 13, and 15 y but not for those aged 7 or 9 y. The prevalence of overweight and obesity increased significantly for all girls combined, with the prevalence of overweight and obesity combined more than doubling among girls aged 11, 13, and 15 y.
View this table:
TABLE 4 . Proportion of boys and girls in each BMI category, including overweight and obese combined (O + O), for the 1985 and 1997 New South Wales data sets
Among boys, the mean weight increased by 3.2 kg, the mean height increased by 2.9 cm, the mean waist girth increased by 0.8 cm, and the mean sum of 3 skinfold thicknesses increased by 5.2 mm for 1985–1997 (Table 3). Among girls, the mean weight increased by 2.7 kg, and the mean height increased by 2.3 cm, but neither the mean waist girth nor the mean sum of 3 skinfold thicknesses increased significantly for the period 1985–1997.
Victoria
The overall student-parent response rate in the Health of Young Victorians Survey, 1997, was 75%. The prevalence of nonoverweight, overweight, obesity, and overweight and obesity combined for the Victorian component of the Australian Health and Fitness Survey, 1985, and the Health of Young Victorians Survey, 1997, is shown (separately for boys and girls aged 7, 8, 9, 10, 11, and 12 y) in Table 5.
View this table:
TABLE 5 . Proportion of boys and girls in each BMI category, including overweight and obese combined (O + O), for the 1985 and 1997 Victorian data sets
Among boys, the proportions who were overweight or obese increased significantly between 1985 and 1997 for those aged 8, 10, and 12 y and all ages combined, but not for those aged 7, 9, or 11 y. Overall, the prevalence of overweight and obesity combined approximately doubled over the period 1985–1997. Among girls, the prevalence of overweight and obesity combined was significantly higher in 1997 than in 1985 for all ages except those aged 8 and 12 y. The prevalence of overweight and obesity combined approximately doubled among girls. Among boys, the mean weight increased by 2.4 kg and the mean height increased by 2.0 cm; for girls the mean weight increased by 2.8 kg and the mean height increased by 2.0 cm over the period 1985–1997 (Table 3).
DISCUSSION
The results for the period 1985–1997 allow confident conclusions. Independent surveys conducted in 3 Australian states all yielded very similar findings: the prevalence of overweight increased by 60–70%, the prevalence of obesity trebled, and the prevalence of overweight and obesity combined doubled. Although we cannot be quite so confident about the findings for the period 1969–1985, the results do indicate that changes in the prevalence of overweight and obesity were far smaller during this 16-y period than they were for the ensuing 12 y. Our findings are generally consistent with studies conducted in the United States (6), Spain (7), and Britain (8).
The mean changes in height and weight are generally consistent with data on the prevalence of overweight and obesity, with weight increasing by 2–3 kg and height increasing by 1–3 cm over the period 1985–1997. That there were some differences between states in the magnitude of these changes was to be expected, given the different age groups involved in the different state surveys. It is notable that although the mean height and weight of SA boys increased between 1969 and 1985, the changes among girls were not statistically significant.
There were also some notable differences between states in the magnitude of changes in mean waist girths and the mean sum of 3 skinfold thicknesses. Although the increase in mean waist girth was substantially greater among SA boys than among NSW boys, increases in mean skinfold thicknesses were of similar magnitudes. Both waist girth and skinfold thicknesses increased substantially among SA girls but not among NSW girls. We do not have an adequate explanation for this but do note that the mean age of SA girls was lower and that in NSW, a greater proportion of girls were past puberty.
Comparing data from independently designed and implemented surveys has both advantages and limitations. The most significant advantage is that findings that are consistent across surveys are likely to represent real phenomena, not spurious or biased results. However, surveys that are not specifically designed for comparison may contain methodologic differences that should be taken into account when forming conclusions. All of the studies we report here used methods of random selection; had adequate-to-high response rates; and used either identical, or at least comparable, methods of measurement. Thus, these 3 key sources of systematic bias are unlikely to have had a significant influence over the results.
Because the 1997 data presented here were all drawn from only 3 of Australia’s 8 states and territories, we cannot be certain that we have identified national trends. However, there are several reasons to believe that other states may not differ greatly. First, for the 1985–1997 period, the findings from the 3 independent state data sets were remarkably consistent, and there is no good reason to believe that these 3 states are similar to each other but different from the remainder of the country. Second, the combined populations of these 3 states represented 67% of the total Australian population in 1997. Third, an earlier study on the point prevalence of overweight and obesity combined among young Australians used the NSW and Victorian data sets employed by the present study in addition to data from a nationally representative sample (5). The point prevalence estimates were similar, showing that these states were not different from the country as a whole. We therefore suggest that the findings presented here probably reflect broader Australian trends.
Our results indicate clearly that from the mid-1980s to the mid-1990s, the prevalence of overweight and obesity combined doubled among young Australians and that the rate of increase over the preceding 16 y was far smaller. The prevalence of overweight and obesity is not only increasing, it is accelerating. These findings should only serve to increase our sense of urgency in identifying and implementing effective responses to this major threat to the public health.
REFERENCES