Trends in the Prevalence of Overweight and Obesity among Young Australians, 1985, 1997, and 2004

Authors


New South Wales Centre for Overweight and Obesity, Level 2, K25-Medical Foundation Building, University of Sydney, New South Wales 2006, Australia. E-mail: Michael_Booth@health.usyd.edu.au

Abstract

Objective: To determine secular trends in overweight/obesity among 7- to 15-year-olds for the periods 1985, 1997, and 2004.

Research Method and Procedures: Data from representative surveys conducted in New South Wales, Australia, in 1985, 1997, and 2004 were analyzed. Height and weight were measured, and BMI categories were created using International Obesity Task Force definitions. Students were grouped as Grades 2 + 4 + 6 and 8 + 10.

Results: The prevalences of overweight/obesity for 1985, 1997, and 2004 were 10.9%, 20.6%, and 25.7% among the younger boys and 10.6%, 19.5%, and 26.1% among the older boys. The average annual rate of increase for 1985 to 1997 was 0.81% and for 1997 to 2004 was 0.73% among the younger boys and was 0.74% and 0.94% for the two periods among the older boys. The prevalences of overweight/obesity in 1985, 1997, and 2004 were 14.0%, 22.0%, and 24.8% among the younger girls and 8.3%, 17.9%, and 19.8% among the older girls, respectively. The average annual rates of increase for the two periods were 0.8% and 0.4% among the younger girls and 0.80% and 0.27% among the older girls. Change in the prevalence of overweight/obesity and socioeconomic status were not associated.

Discussion: Over the period 1985 to 1997, the prevalence of overweight and obesity increased significantly among the younger and older boys and the younger girls. The prevalence of overweight, but not obesity, increased among the older girls over this period. Over the period 1997 to 2004, the prevalence of overweight/obesity combined increased significantly among boys of both age groups but not among girls.

Introduction

The current study reports on changes in the prevalence of overweight and obesity among 5- to 15-year-old students living in Australia's most populated state [New South Wales (NSW)]1 based on representative population surveys conducted in 1985, 1997, and 2004. The associations between changes in the prevalence of overweight and obesity and socioeconomic status (SES) are also examined. NSW accommodates 33% of Australia's population, and its demographic structure is similar to those of most developed economies, except for its wide cultural diversity (∼21% of the population were born outside Australia, and 24% speak a language other than English at home) (1).

Research Methods and Procedures

1985 Data Collection: The 1985 Australian Health and Fitness Survey (AHFS)

The methods of the AHFS 1985 have been reported in detail elsewhere (2). Briefly, the AHFS 1985, was a large-scale national study (n = 8484) that assessed many aspects of health-related fitness, including serum analysis, among students 7 to 15 years old. A random selection of schools was stratified by state, and students were selected at random within schools. Data for NSW (N = 1227) were extracted to allow comparison with the 1997 and 2004 data.

1997 Data Collection: The NSW Schools Fitness and Physical Activity Survey (NSWSFPAS) 1997

The methods employed in the NSWSFPAS 1997 have been described in detail elsewhere (3). The study was approved by the University of Sydney Human Research Ethics Committee, by the NSW Department of Education, and by the NSW Catholic Education Commission. NSWSFPAS 1997 was a representative population survey of students attending Grades 2, 4, 6, 8, and 10 in primary and secondary schools in rural and urban NSW (N = 5518). Primary and secondary schools were selected at random from each education sector proportional to enrollment in that sector, and then one class in each target grade was selected at random within each school. Fundamental movement skill proficiency and a wide range of health-related fitness variables were assessed, and comprehensive self-reported information on participation in and perceptions of physical activity was collected from students in Grades 8 and 10. The data were collected during February and March 1997.

2004 Data Collection: The NSW Schools Physical Activity and Nutrition Survey (SPANS) 2004

The methods employed in SPANS 2004 (N = 5402) have been described in detail elsewhere (4). The study was approved by the University of Sydney Human Research Ethics Committee, by the NSW Department of Education and Training, and by the NSW Catholic Education Commission. The methods of this survey were virtually identical to those of the NSWSFPAS 1997, although some self-report measures were different, and blood samples were collected from Grade 10 students in 2004. The data were collected during February, March, and April 2004.

For all three surveys: Students’ school grade, sex, date of birth, and postal code of residence were recorded; height and weight were directly measured with shoes and socks removed and with their pockets emptied. Height and weight values were used to calculate BMI (kilograms per meter squared), and the International Obesity Task Force definitions were used to categorize students as healthy, overweight, or obese (5). Postal code of residence was used as a proxy measure of SES. The Australian Bureau of Statistics calculated values of the Index of Relative Socioeconomic Disadvantage based on data collected in the population census most proximal to the survey date, and these were used to rank students into tertiles of SES (6).

Statistical Analyses

Analyses were conducted using version 9.1 of the SAS System for Windows (SAS Institute, Inc., Cary, NC). The number and proportion of students in each BMI category were tabulated for each sex and school year group by survey year, with results additionally summarized by SES tertile. Changes in prevalence over time were assessed using the survey logistic regression procedure to account for the clustered nature of the surveys. Equality in the rates of change over time was assessed within the SURVEYLOGISTIC procedure (SAS Institute, Inc.). The difference between the 1997 and 2004 time parameters was constructed using a linear contrast allowing for the different lengths of time of the two periods. The linear contrast was then tested against 0 (indicating equality of the rates of change).

Results

Participation Rates and Characteristics of the Samples

The participation rates of the three surveys varied across boys and girls in each school grade. For the AHFS 1985, the participation rates ranged from 56% to 72%, the participation rates for the NSWSFPAS 1997 ranged from 71% to 92%, and for SPANS 2004, the participation rates ranged from 49% to 77%. The mean ages for students in Grades 2, 4, 6, 8, and 10 for the AHFS were 7.5, 9.5, 11.5, 13.5, and 15.5 years, respectively, and for both the NSWSFPAS 1997 and the SPANS 2004 were 7.3, 9.3, 11.3, 13.3, and 15.3 years, respectively.

Secular Trends

Figure 1 (A and B) and Table 1 show the prevalence of overweight and obesity (separately and combined) stratified by sex for younger (Grades 2, 4, and 6) and older (Grades 8 and 10) children in 1985, 1997, and 2004. The table also shows, for each demographic stratum, changes in prevalence, the associated odds ratio, and the average annual rate of change in prevalence. Figure 2 (A and B) and Table 2 show the prevalence of overweight and obesity combined stratified by sex, age group, and tertiles of SES for 1985, 1997, and 2004. Table 2 also shows, for each SES tertile, changes in prevalence, the associated odds ratio, and the average annual rate of change in prevalence.

Figure 1.

(A) Prevalence of overweight (Ow), obesity (Ob), and overweight/obesity combined (OwOb) among Grades 2, 4, and 6 boys (2,4,6) and among Grades 8 and 10 boys (8,10) in 1985, 1997, and 2004. (B) Prevalence of overweight (Ow), obesity (Ob), and overweight/obesity combined (OwOb) among Grades 2, 4, and 6 girls (2,4,6) and among Grades 8 and 10 girls (8,10) in 1985, 1997, and 2004.

Table 1.  Prevalences of overweight, obesity, and overweight and obesity combined for boys and girls in Grades 2, 4, and 6 combined and Grades 8 and 10 combined. Prevalences in 1985, 1997, and 2004 are shown with the differences between 1985 and 1997 and between 1997 and 2004. Odds ratios, 95% confidence intervals, and comparisons between surveys are also shown. Annual rate of change in prevalence (pa) between 1985 and 1997 and between 1997 and 2004 is also shown for overweight, obesity, and overweight and obesity combined for both grade groupings for boys and girls separately
 Boys
 19851997Diff.Odds ratio 95% confidence intervalRate of change pa (%)19972004Diff.Odds ratio 95% confidence intervalRate of change pa (%)Equal rates of change (P-value)
  • *

    p < 0.05;

  • p < 0.01;

  • p < 0.001.

  • §

    Lower confidence limit constrained to be positive.

Grades 2, 4, 6           
Overweight9.0 (6.4–11.5)15.0 (13.1–17.0)6.1 (2.9–9.3)1.8 (1.28, 2.52)0.5115.0 (13.1–17.0)18.5 (16.7–20.4)3.5 (0.8–6.2)1.3 (1.06, 1.58)*0.500.6
Obese2.0 (1.0–3.0)5.6 (4.4–6.8)3.6 (2.0–5.2)2.9 (1.65, 5.20)0.305.6 (4.4–6.8)7.2 (5.9–8.4)1.6 (−0.2–3.3)1.3 (0.96, 1.75)0.220.2
Combined10.9 (8.5–13.4)20.6 (18.3–23.0)9.7 (6.3–13.0)2.1 (1.59, 2.80)0.8120.6 (18.3–23.0)25.7 (23.4–28.0)5.1 (1.8–8.4)1.3 (1.11, 1.61)0.720.3
Grades 8, 10           
Overweight10.2 (6.8–13.6)15.6 (13.0–18.2)5.4 (1.2–9.7)1.6 (1.08, 2.47)*0.4515.6 (13.0–18.2)19.4 (16.1–22.7)3.8 (−0.4–8.0)1.3 (0.98, 1.73)0.540.9
Obese0.4 (0.0§–1.3)3.9 (2.3–5.5)3.4 (1.6–5.3)9.0 (1.26, 63.67)*0.293.9 (2.3–5.5)6.7 (4.5–8.8)2.8 (0.1–5.5)1.8 (1.05, 3.13)*0.400.3
Combined10.6 (7.1–14.2)19.5 (16.0–23.0)8.9 (3.9–13.8)2.0 (1.33, 3.11)0.7419.5 (16.0–23.0)26.1 (21.5–30.7)6.6 (0.8–12.4)1.5 (1.06, 2.02)*0.940.9
 Girls
Grades 2, 4, 6           
Overweight11.9 (8.5–15.3)16.6 (14.4–18.7)4.6 (0.6–8.7)1.5 (1.03, 2.09)*0.3916.6 (14.4–18.7)17.6 (15.7–19.4)1.0 (−1.8–3.8)1.1 (0.88, 1.31)0.140.4
Obese2.0 (0.7–3.3)5.4 (4.0–6.8)3.4 (1.4–5.3)2.7 (1.35, 5.52)0.285.4 (4.0–6.8)7.2 (5.9–8.5)1.8 (−0.1–3.7)1.4 (0.97, 1.90)0.260.4
Combined14.0 (9.7–18.2)22.0 (19.5–24.4)8.0 (3.1–12.9)1.7 (1.19, 2.52)0.6722.0 (19.5–24.4)24.8 (22.5–27.1)2.8 (−0.6–6.2)1.2 (0.97, 1.41)0.400.3
Grades 8, 10           
Overweight7.8 (2.8–12.9)15.2 (12.3–18.1)7.4 (1.6–13.2)2.1 (1.04, 4.37)*0.6215.2 (12.3–18.1)16.2 (13.0–19.4)1.0 (−3.3–5.3)1.1 (0.78, 1.50)0.140.2
Obese0.5 (0.0§–1.4)2.7 (0.8–4.6)2.2 (0.1–4.4)5.6 (0.74, 42.60)0.192.7 (0.8–4.6)3.6 (2.3–4.9)0.9 (−1.4–3.2)1.3 (0.61, 2.97)0.120.4
Combined8.3 (3.5–13.1)17.9 (14.5–21.4)9.6 (3.7–15.6)2.4 (1.26, 4.66)0.8017.9 (14.5–21.4)19.8 (16.3–23.3)1.8 (−3.1–6.8)1.1 (0.83, 1.57)0.260.2
Figure 2.

(A) Prevalence of overweight/obesity combined in low, medium, and high SES tertiles among Grades 2, 4, and 6 boys (2,4,6) and among Grades 8 and 10 boys (8,10) in 1985, 1997, and 2004. (B) Prevalence of overweight/obesity combined in low, medium, and high SES tertiles among Grades 2, 4, and 6 girls (2,4,6) and among Grades 8 and 10 (8,10) girls in 1985, 1997, and 2004.

Table 2.  Prevalences of overweight and obesity combined for boys and girls in Grades 2, 4, and 6 combined and Grades 8 and 10 combined, stratified by tertiles of socioeconomic status. Prevalences in 1985, 1997, and 2004 are shown with the differences between 1985 and 1997 and between 1997 and 2004. Odds ratios, 95% confidence intervals, and comparisons between surveys are also shown. Annual rate of change in prevalence (pa) between 1985 and 1997 and between 1997 and 2004 is also shown for overweight and obesity combined for both grade groupings for boys and girls separately
 Boys
 19851997Diff.Odds ratio 95% confidence intervalRate of change pa (%)19972004Diff.Odds ratio 95% confidence intervalRate of change pa (%)Equal rates of change (P-value)
  • *

    p < 0.05;

  • p < 0.01;

  • p < 0.001.

  • §

    Lower confidence limit constrained to be positive.

Grades 2, 4, 6           
Low11.0 (6.2–15.8)19.5 (14.8–24.2)8.5 (1.8–15.2)2.0 (1.2, 3.4)0.7119.5 (14.8–24.2)24.9 (20.5–29.4)5.4 (−1.0–11.9)1.4 (0.9, 2.0)0.780.6
Medium13.7 (7.0–20.4)23.7 (19.6–27.7)10.0 (2.1–17.8)2.1 (1.2, 4.0)*0.8323.7 (19.6–27.7)27.0 (23.1–30.8)3.3 (−2.2–8.9)1.2 (0.9, 1.6)0.470.3
High9.1 (1.0–17.2)18.8 (16.1–21.5)9.7 (1.2–18.2)2.3 (0.9, 5.8)0.8118.8 (16.1–21.5)25.2 (21.4–29.0)6.4 (1.7–11.1)1.5 (1.1, 1.9)0.910.6
Grades 8, 10           
Low13.6 (2.5–24.6)20.7 (16.0–25.3)7.1 (−4.9–19.1)1.6 (0.7, 4.0)0.5920.7 (16.0–25.3)28.7 (22.4–34.9)8.0 (0.2–15.8)1.5 (1.0, 2.3)1.140.7
Medium7.2 (3.1–11.4)18.2 (12.0–24.5)11.0 (3.5–18.5)2.8 (1.4, 5.6)0.9218.2 (12.0–24.5)26.2 (21.5–30.9)8.0 (0.2–15.8)1.6 (1.0, 15.8)1.140.8
High15.5 (6.6–24.4)19.8 (13.2–26.5)4.3 (−6.8–15.4)1.3 (0.6, 2.9)0.3619.8 (13.2–26.5)23.3 (16.3–30.2)3.4 (−6.2–13.1)1.2 (0.7, 2.2)0.490.9
 Girls
Grades 2, 4, 6           
Low18.9 (13.0–24.9)25.1 (21.2–29.1)6.2 (−0.9–13.3)1.4 (0.9, 2.1)0.5225.1 (21.2–29.1)27.0 (23.0–31.0)1.8 (−3.8–7.4)1.1 (0.8, 1.5)0.260.6
Medium10.4 (3.3–17.5)23.1 (19.0–27.2)12.7 (4.5–20.9)2.4 (1.1, 5.1)*1.0623.1 (19.0–27.2)23.6 (19.7–27.6)0.6 (−5.1–6.2)1.0 (0.8, 1.4)0.080.08
High13.8 (4.8–22.7)18.1 (14.1–22.1)4.4 (−5.4–14.1)1.5 (0.7, 3.3)0.3618.1 (14.1–22.1)23.6 (20.1–27.0)5.5 (0.2–10.7)1.4 (1.0, 1.9)0.780.7
Grades 8, 10           
Low7.4 (1.6–13.3)19.6 (13.9–25.4)12.2 (4.0–20.4)3.0 (1.3, 7.1)1.0219.6 (13.9–25.4)24.9 (17.8–32.0)5.3 (−3.9–14.4)1.4 (0.8, 2.3)0.750.4
Medium10.5 (2.9–18.1)20.1 (15.3–24.9)9.6 (0.6–18.6)2.1 (0.9, 4.9)0.8020.1 (15.3–24.9)15.8 (10.2–21.3)−4.3 (−11.7–3.0)0.7 (0.4, 1.2)−0.620.06
High6.7 (0.0†–14.4)14.7 (10.2–19.2)8.0 (−0.9–17.0)2.6 (0.7, 9.0)0.6714.7 (10.2–19.2)18.3 (13.9–22.6)3.6 (−2.7–9.8)1.3 (0.8, 2.1)0.510.6

1985 to 1997

Among boys, the patterns of change were very similar for the older and younger groups. Between 1985 and 1997, the prevalence of overweight, obesity, and overweight and obesity combined increased by ∼0.5%, 0.3%, and 0.8% per annum, respectively. The changes in the prevalence of overweight, obesity, and overweight/obesity combined between 1985 and 1997 were all statistically significant for both the younger and older boys.

Among girls, the prevalence of overweight, obesity, and overweight/obesity combined increased ∼0.4%, 0.3%, and 0.7%, respectively, among the younger girls and by 0.6%, 0.2%, and 0.8%, respectively, among the older girls. The prevalence of all three BMI categories increased significantly among the younger and older girls with the exception of the obese category among the older girls.

With regard to change stratified by SES tertiles among boys, there was no consistent pattern of change. Among the younger boys, all three tertiles changed by about the same magnitude, with the changes being statistically significant for the low and medium tertiles. Among the older boys, the greatest increase occurred in the middle tertile (the only statistically significant difference). Among the younger girls, the greatest change occurred in the middle SES tertile (the only statistically significant difference), and among the older girls, the greatest change was in the low SES tertile (the only statistically significant difference), with little difference between the middle and high SES tertiles.

1997 to 2004

Among the younger boys, the prevalence of overweight, obesity, and overweight/obesity combined increased ∼0.5%, 0.2%, and 0.7%, respectively, and among the older boys by 0.5%, 0.4%, and 0.9%, respectively. The prevalence of overweight and overweight/obesity combined increased significantly among the younger boys, and the prevalence of obesity and overweight/obesity combined increased significantly among the older boys over this period. The average annual rates of change over the two periods were very similar among the younger boys but were somewhat greater over the latter period among the older boys, suggesting that the incidence of overweight and obesity may be increasing over time among the older group. However, none of the differences in the rates of range between the two periods were statistically significant for either group of boys.

Among the younger girls, the prevalence of overweight, obesity, and overweight/obesity combined increased by ∼0.14%, 0.3%, and 0.4%, respectively, and among the older girls by 0.14%, 0.12%, and 0.3%, respectively. None of the changes in prevalence between 1997 and 2004 were statistically significant among girls. Although the average annual rates of change were smaller over the period 1997 to 2004 than over the period 1985 to 1997, none of the differences between the rates of change were statistically significant.

With regard to change stratified by SES tertiles during the period 1997 to 2004, there were no consistent associations. Among the younger boys, all three tertiles changed by similar amounts, with the largest increase occurring in the high SES tertile (the only one that was statistically significant). Among the older boys, the prevalence of overweight/obesity increased equally in the low and middle tertiles (both p = 0.04). Among the younger girls, the largest increase occurred in the high SES tertile (p = 0.05), and there was no consistent difference across tertiles among the older girls.

Discussion

This study found that the prevalence of overweight/obesity in 2004 among boys and girls 7 to 15 years old was ∼25%. The prevalence increased by 0.7% to 0.8% per year among the four age/sex groups, over the period 1985 to 1997, but the secular trends of separate age/sex groups seem to have diverged over the period 1997 to 2004. Among girls, the average annual rate of increase in the prevalence of overweight/obesity during 1997 to 2004 was approximately one-half that of the period 1985 to 1997, although these differences were not statistically significant. Among boys, the average annual rate of increase in the prevalence of overweight and obesity did not change substantially among the younger boys but increased by almost 30% among the older boys. Although the differences between the average annual rates of change in prevalence were not statistically significant, it is notable that the prevalence of overweight and obesity was statistically greater in 2004 compared with 1997 for both groups of boys but for neither group of girls. This pattern of results suggests that the secular trends in the prevalence of overweight and obesity may be diverging for boys and girls, but the evidence is not sufficiently strong to support a firm conclusion.

This study failed to find any consistent associations between SES and secular trends in the prevalence of overweight/obesity among either boys or girls, at least in this region of Australia. It should be noted, however, that postal code of residence is not a very sensitive indicator of SES.

It is encouraging to see that the rate of increase appears to be slowing among girls, but another survey (in ∼2010) is required to confirm that this is a true trend change and not simply a minor fluctuation. In addition, it should be kept in mind that although the rate of increase may be decelerating, it is still positive and that more and more young Australians are becoming overweight or obese. If these trends are confirmed, the findings suggest that different intervention models may need to be applied to boys and girls and that both qualitative and quantitative research be undertaken to understand why the divergence in secular trends may be occurring and what factors may be influencing the behaviors of older boys.

Acknowledgments

The NSWSFPAS was funded by the NSW Department of Education. SPANS 2004 was funded by NSW Health and the NSW Department of Education and Training. M.B. is supported by a National Health and Medical Research Council research fellowship.

Footnotes

  • 1

    Nonstandard abbreviations: NSW, New South Wales; SES, socioeconomic status; AHFS, Australian Health and Fitness Survey; NSWSFPAS, NSW Schools Fitness and Physical Activity Survey; SPANS, Schools Physical Activity and Nutrition Survey.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Ancillary