Age, period and birth cohort effects on the prevalence of overweight and obesity among Taiwanese adolescents: a national population-based study

Age, period and birth cohort effects on the prevalence of overweight and obesity among Taiwanese... Abstract Background The age–period–cohort effects on youth overweight and obesity among junior and senior high school students in Taiwan is not clear. Methods We utilized the 2006–14 nationally representative School Physical Fitness Dataset. Based on the International Obesity Task Force cut-off criteria, a log-linear age–period–cohort analysis was performed to determine the influence of age, period and cohort on the trend in being overweight, obese and severely obese for both sexes. Results The final dataset included 1073173 individuals (n = 520 382 boys and 552 791 girls). For girls, the prevalence of overweight and obesity declined with age, and the prevalence of overweight declined over time. For boys, the prevalence of overweight and obesity declined with age and over time from 2006 to 2014. The prevalence of severe obesity declined over time and increased with age for the boys. The younger birth cohorts had greater odds of being overweight, obese and severely obese than the older birth cohorts. Conclusions After differentiating the age–period–cohort effects, the data suggested a decreasing temporal trend in overweight and obesity among adolescents in Taiwan from 2006 to 2014. Among the birth cohorts of the 1990s, the younger cohorts had greater odds of being overweight and obese than the older cohorts when they reached adolescence. adolescents, age–period–cohort study, body mass index, epidemiology, growth, severe obesity Introduction Overweight and obesity (defined as abnormal or excessive fat accumulation, respectively, with negative impacts on health) has reached epidemic levels in many parts of the world.1,2 Globally, overweight and obesity affected an estimated 23.8% of boys and 22.6% of girls in developed countries in 2013.2 From a public health perspective, tracking the trends in overweight and obesity among youth is important. Excessive weight among youths is associated with medical consequences3,4 and significant financial healthcare costs.5 Furthermore, childhood obesity may increase the susceptibility of children to becoming overweight or obese adults, which consequently leads to more healthcare costs.6,7 Compared with the drastically increasing trend in childhood obesity over the past four decades, emerging data in recent years have suggested that the increase in the prevalence of childhood obesity may have stopped in developed countries.8–12 In the US, the prevalence of childhood obesity remained at ~15–17% from 2005 to 2014.11 UK data showed a similar phenomenon from 2004 to 2013.12 This trend may be related to less obesogenic environments,8 potential method biases, or subgroup heterogeneity.9,13 In Taiwan, some research has examined the prevalence of overweight and obesity in children and adolescents. For instance, 2 and 1% increases among 12–15-year-old boys and girls, respectively, were documented between 1980 and 1994.14 The findings also indicated that the prevalence of overweight and obesity increased from 19.8 to 26.8% and 15.2 to 16.5% among 6–18-year-old boys and girls, respectively, from 1999 to 2001.15 Overall, these statistics imply that the prevalence of overweight and obesity has increased among youths in Taiwan over the past few decades. However, because public attention and policies targeting childhood obesity prevention and control have increased recently,16–18 the prevalence of childhood obesity in Taiwan may plateau, as has been observed in advanced countries.8–12 Importantly, the temporal trend in prevalence may be confounded by varying age distributions and structures of birth cohorts across years. Although there have been reports of temporal trends in the prevalence of overweight and obesity in Taiwan,14,15 these studies are either based on data with a relatively small sample size14 or have not examined the period or birth cohort effects in addition to the age group effects.15 To the best of our knowledge, few studies have discriminated the effects of age, period and cohort on the prevalence of overweight and obesity among the youth population in Taiwan. To untangle the inter-independent problem between these three variables, previous research has utilized the age–period–cohort (APC) analysis model to successfully explore the prevalence of overweight and obesity in adult populations.19,20 Therefore, based on the national School Physical Fitness Database (SPFD) of Taiwan, the aim of the current study was to advance our knowledge on the temporal trend in the prevalence of overweight and obesity in adolescents in Taiwan by examining the influence of age, period and cohort simultaneously. Methods Study population This study was based on the SPFD, which was established by the Taiwan Ministry of Education (MOE) in 2006. The SPFD contains detailed information on the long-term records of physical fitness components of students, such as the body mass index (BMI), cardiovascular fitness, flexibility, and muscular strength and endurance. Physical fitness is measured by physical educational teachers and school nurses who are qualified to perform these measurements following a detailed standardized instruction manual. In addition to fitness information, other demographic data, including the age, sex and grade of each student, are recorded in this dataset. Because uploading fitness data from every student into the SPFD is compulsory according to government laws and regulations,21 the SPFD is representative of the national youth population aged 10–18 years in Taiwan. The original dataset included 1 369 216 height and weight observations. After excluding the observations with missing body weight, height, or birth date values (n = 16 179, 1.18%), unreasonable biological height, weight and/or age values (i.e. age < 60 months, age > 228 months, weight < 10 kg, weight > 200 kg, height < 50 cm or height > 300 cm; n = 3124, 0.23%) and extreme BMI values (>+5 SD or <−5 SD; n = 532, 0.04%), 1 342 282 valid observations remained. Because the majority of the observations were obtained from 2006 to 2014, this study focused on the trends of weight status during this period. Adolescents who were 12–18 years old at the time of the physical examination from 2006 to 2014 were included in the analysis, because this age range contained data collected over a longer period. These children were born between 1994 and 1997. The final sample size was 520 382 boys and 552 791 girls. This study was approved by the local Institutional Ethics Committee at National Taiwan Normal University. Measures The body weights and heights of the children were measured at school by the school nurses, and the BMIs of the adolescents were calculated as the weight (kg)/height2 (m2). The weight statuses were defined using the sex- and age-specific cut-off points from the International Obesity Task Force (IOTF),22 which are recommended for use in international comparisons of the prevalence of overweight and obesity. Based on the IOTF cut-off BMI points, the weight statuses were categorized as follows: overweight, obese and severely obese corresponding to an 18-year-old adult BMI of 25, 30 and 35, respectively. For instance, the international cut-off point for overweight is 21.22 and 21.68 for 12-year-old boys and girls, respectively. In the current study, the obese category encompassed individuals who were obese or severely obese, and the overweight category encompassed individuals who were overweight, obese or severely obese. Statistical analyses The proportion of weight statuses by year of examination and the children’s ages were used to estimate the prevalence of overweight, obesity and severe obesity. Using the Statistical Analysis Software (SAS) GENMOD procedure (SAS 9.3, SAS Institute Inc., Cary, NC, USA) (specifying log link function and Poisson count distribution offset by the denominator population), the Poisson log-linear APC model was applied to analyze binned data of the yearly prevalence. The full model specified all ages, periods (reference = exam years 2009 and 2010) and cohorts (reference = birth year 1995) without an intercept. The deviance difference between the full model and the model omitting the factor of interest was determined to test the effect of each factor on the prevalence by comparing the data with the chi-square distribution to obtain a P-value. The models estimated the prevalence rate ratios for the period (exam year) and cohort (birth year). The prevalence of weight status by age with adjustments for the period and cohort effects was estimated using the least square mean approach. Results Table 1 presents the distributions of overweight, obesity and severe obesity among adolescents aged 12–18 years in Taiwan from 2006 to 2014. The population corresponds to children born from 1994 to 1997. Generally, the prevalence of overweight and obesity increased with the exam year when age was fixed (column-wise trend in Table 1). Additionally, the prevalence of overweight and obesity decreased with age (row-wise trend in Table 1). Table 1 Trends in the prevalence of adolescent overweight, obesity and severe obesity in Taiwan Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Table 1 Trends in the prevalence of adolescent overweight, obesity and severe obesity in Taiwan Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Table 2 presents whether age, period (exam years) and/or cohort (birth years) affect the prevalence of overweight, obesity and severe obesity. The global tests showed the significance of the age, period and cohort effects. The direction of the temporal effect was revealed by delineation of the prevalence by age, time and cohort when the other two temporal factors were controlled, as shown in Figs 1 and 2. The findings are described by the age, period and cohort effects as follows. Table 2 Testing the age, period and cohort effects for the three weight statuses stratified by sex Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Note: The individual effect of age, period or cohort was assessed by comparing the difference in deviance between the full three-factor model (age–period–cohort) and the two-factor model that omitted the factor of interest. Table 2 Testing the age, period and cohort effects for the three weight statuses stratified by sex Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Note: The individual effect of age, period or cohort was assessed by comparing the difference in deviance between the full three-factor model (age–period–cohort) and the two-factor model that omitted the factor of interest. Fig. 1 View largeDownload slide Estimated age, period and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent girls: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends show risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 1 View largeDownload slide Estimated age, period and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent girls: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends show risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 2 View largeDownload slide Estimated age, period, and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent boys: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends are risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 2 View largeDownload slide Estimated age, period, and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent boys: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends are risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Age effect: The age effect significantly affected the prevalence of overweight and obesity for both girls and boys. According to Table 1 and Figs 1 and 2, the prevalence of overweight and obesity decreased with age in the adolescent population (both sexes). The age trend for severe obesity was significant only for boys. Period effect: Regarding the period effect, the prevalence of overweight, obesity and severe obesity for boys decreased over time from 2006 to 2014. For girls, only the prevalence of overweight varied by exam year. The prevalence of overweight in girls was increased by 1.12-fold (95% confidence interval [CI]: 1.01–1.24) and 1.13-fold (95% CI: 1.05–1.20) in 2006 and 2007, respectively, when compared with the prevalence of overweight from 2009 to 2010. For boys, the prevalence rates of overweight, obesity, and severe obesity all varied significantly by exam year (Table 2). Compared with the prevalence of overweight from 2009 to 2010, the prevalence of overweight in 2006 was increased by 1.18-fold (95% CI: 1.08–1.28), and the prevalence of overweight in 2014 was reduced by 0.88-fold (95% CI: 0.80–0.97). Compared with the prevalence of obesity in boys from 2009 to 2010, the prevalence of obesity in 2006 was increased by 1.45-fold (95% CI: 1.23–1.71), and the prevalence in 2014 was reduced by 0.81-fold (95% CI: 0.67–0.98). Similarly, compared with the prevalence of severe obesity in boys from 2009 to 2010, the prevalence of severe obesity in 2006 was increased by 1.69-fold (95% CI: 1.07–2.66), and the prevalence in 2014 was reduced by 0.62-fold (95% CI: 0.38–1.01). Generally, the results suggested a decreasing prevalence of overweight, obesity and severe obesity among boys. Cohort effect: The prevalence of overweight, obesity, and severe obesity varied significantly by birth year. The prevalence rates of overweight and obesity were greater for children born in the later years than those for children born in the earlier years (Figs 1 and 2). Compared with boys born in 1994, boys born in 1997 had 1.35 (95% CI: 1.27–1.44) times, 2.01 (95% CI: 1.78–2.28) times and 2.91 (95% CI: 2.09–4.05) times higher prevalence rates of overweight, obesity, and severe obesity, respectively. Compared with girls born in 1994, girls born in 1997 had 1.15 (95% CI: 1.07–1.25) times and 1.25 (95% CI: 1.05–1.48) times higher prevalence rates of overweight and obesity, respectively. Discussion Main findings of this study This study examined eight years of national data from the SPFD (2006–14) for children and adolescents in Taiwan and provided unique comparative information regarding trends in the prevalence of preadolescent and adolescent overweight, obesity and severe obesity (Tables 1 and 2). The prevalence of overweight and obesity among adolescents has stabilized or even decreased during the last decade. The prevalence of overweight and obesity decreased with age among Taiwanese adolescents, but the prevalence of severe obesity increased with age. Interestingly, the younger birth cohorts were more likely to be overweight or obese than the older birth cohorts. What is already known on this topic Regardless of the definition used for overweight or obesity, data from 1990 to 2000s indicated that the prevalence was increasing in Taiwan. In Taipei (the capital city of Taiwan), the prevalence of obesity among 12–15-year-old school boys significantly increased from 12.4 to 16.4% from 1980 to 1994, respectively.14 A similar increasing trend in the prevalence of overweight and obesity was reported among students aged 6–18 years from 1999 to 2001.15 Based on the representative data from all Taiwanese adolescents (n = 38 921), the prevalence of overweight increased from 14.1 to 18.6% and 12.8 to 13.0% in boys and girls, respectively.15 The prevalence of obesity also increased from 5.7 to 8.2% and 2.4 to 3.6% in boys and girls, respectively.15 What this study adds Similar to data from economically advanced countries or areas,8–10 such as the USA11 and UK,12 the prevalence of childhood obesity in Taiwan appeared to stop increasing. From 2006 to 2014, the prevalence of overweight and obesity decreased among adolescents aged 12–18 years. Although the exact explanation for this recent decline is unclear, policy measures to prevent and control the obesity epidemic have been in place since the mid-2000s. In 2005, the Ministry of Education issued an executive order that banned the sale of sweetened beverages and unhealthy snacks in elementary and junior high schools. Additionally, the Health-Promoting School Program, which includes promotion of a healthy weight status by improving school lunch quality, offering nutrition education, and encouraging children’s healthy lifestyle (e.g. sleep, balance diet and active life), was completely implemented in all elementary and junior high schools in Taiwan by 2008.23 In 2013, the Taiwanese Food and Drug Administration has implemented the regulations regarding to the advertisement of foods that ‘are unsuitable for long-term consumption for children’ to children.17 These policies could have created an environment that counteracted the rising trend in overweight and obesity in adolescents. Further research is warranted to elucidate the impacts of these policies. The age trend in the prevalence of overweight and obesity can reveal the age trajectory of the weight status during life stages. Our current results indicate inverse relationships between overweight and obese statuses and age. The finding agrees with earlier studies from Taiwan14,15,18,24 and China24,25 that found that younger children were at greater risk of being overweight and obese than older children. For example, the prevalence of overweight/obesity among boys declined from 27.6 to 20.8%, and a similar decrease in the prevalence of overweight/obesity from 17.1 to 8.7% was observed among girls.15 The reduction in the prevalence with increased age might reflect the increased weight-related concerns or body dissatisfaction among older adolescents,26 which might subsequently lead to an increase in the practice of weight-loss behaviors.26 However, the decreased prevalence of overweight and obesity with age in Taiwan and China also raises a question concerning the application of the international growth references for these Asian populations. For instance, the prevalence of overweight and obesity that appears to be decreasing with age could result from slowing of the weights and heights of the target populations during adolescence when compared with the reference population. Although the IOTF reference incorporates data from Hong Kong, further research is needed to examine the fitness of international growth references to these Asian populations. Another finding that deserves attention is that the prevalence of severe obesity increases with age. This finding implies that there could be different groups of adolescents with diverse growth trajectories during this life stage. Although the majority of adolescents had lower risks of overweight and obesity while growing up, a small proportion might need additional medical care to prevent the severe obesity status from carrying over to adulthood. Relative to the birth cohort born during 1994, the cohort born toward the end of 1997 tended to experience increased odds of being overweight and obese. Another study on Chinese adults also showed that the younger cohort exhibited significantly increased BMI values compared to the older cohort.27 Birth cohort effects reflect environmental changes, such as maternal increased food consumption and sedentary behaviors, during the economic transition. The per capita gross domestic product increased from $8 132 in 1990 to $14 519 in 2000. This economic growth in Taiwan over recent decades might have resulted in an ‘obesogenic’ environment, with a lower cost of energy-dense foods and more convenient lifestyle,19 since the prevalence of obesity was rising during this period as mentioned above. Maternal obesity and diet quality might affect the offspring’s future risks of obesity.28–30 The modified risks of obesity due to the intrauterine environment might explain the observed cohort differences in obesity. The younger birth cohorts’ greater prevalence of overweight and obesity might reflect their exposure to the increasingly obesity-promoting environment when they were fetuses during the 1990s. Limitations of this study Some limitations of the current study should be considered. First, we were unable to examine township and city district-specific trends in overweight, obesity and severe obesity. Prior research has suggested that the prevalence of obesity is more severe in poor and rural areas than that in the more urbanized areas of Taiwan.23 Additionally, interventional strategies against overweight and obesity might be developed and implemented at the regional level. Understanding which of these policies are most effective is not possible by examining the national data used in this study. Including geographic information in future research will increase our understanding of the geographic prevalence of overweight and obesity in Taiwan and the effectiveness of interventional strategies. Another limitation was the lack of comprehensive information for the students in the current study (related to family and ethnic groups). Some research has suggested that the prevalence may be influenced by ethnic groups.31,32 Because 92.7% of Taiwanese are of the Han nationality, the current findings provide a general picture of the effects of the age, period and birth cohort on the prevalence of overweight and obesity in Taiwan. Third, the criteria for overweight, obesity and severe obesity in the current study were defined based on the BMI values, which is a popular and straightforward measurement method designed for epidemic research. However, the use of an adiposity measure, such as skinfold thickness, should be considered in future studies. Fourth, the data were originally intended for administrative purposes, and the measurements were taken by school personnel. Despite the standardized protocol, measurement errors could still have occurred. However, the measurement errors should be random and canceled out when the data for the whole population are pooled. Moreover, there was no evidence of population-level systematic errors in the body weight and height measurements. Fifth, the ages of the children in this study spanned from junior to senior high school. Compulsory education in Taiwan extends for 9 years until junior high school. Thus, not all junior high school students go on to senior high school. Nevertheless, the enrollment rates were high for the junior and senior high schools. For example, in 2008, 97% of children aged 12–14 years were enrolled in junior high school, and 91% of children aged 15–17 years were enrolled in senior high school. In 2014, 98% of 12–14-year-old children were enrolled in junior high school, and 94% of 15–17-year-old children were enrolled in senior high school. The small differences in the enrollment rates between junior and senior high school cannot truly explain the age trend in overweight and obesity. Despite these limitations, our current study possesses several strengths, such as the large nationally representative dataset, which is free from self-report bias through standardized collection protocols for the anthropometric measurements of height and weight. More importantly, the current study utilized a refined statistical approach (the APC analysis) to overcome the inherent problem of discriminating the effects of age, period and birth cohort and enabled us to demonstrate the independent effects of age, period and birth cohort on the prevalence of overweight and obesity among Taiwanese youths. In conclusion, to the best of our knowledge, the prevalence of overweight and obesity among Taiwanese youth has stopped increasing and even started declining. Additionally, the prevalence of overweight and obesity decreased significantly as the children aged. Nevertheless, the prevalence of severe obesity increased with age. Apart from the contribution of age, clear cohort differences were present for overweight and obesity among Taiwanese youths. As noted, the cohort differences may reflect differences in the years of exposure to increasing environmental obesogenecity during the 1990s among these birth cohorts. Although the prevalence in Taiwan did not increase from 2006 to 2014, continual monitoring of the obesogenic environment and weight status of children is necessary to understand the influence of policies on temporal trends. Conflicts of interest The authors declared no conflicts of interests. Acknowledgements We thank Drs Chung-Ching Chen and Tzyy-Yuang Shiang for preparing the article. Authors’ contributions YKC and HJC conceived the study and designed the analysis processes. CHC and HJC were responsible for the data cleaning, analysis and interpretation. YKC, HJC and CHC drafted sections of the article. All authors were involved in manuscript review and revision and approved the final version of this article. Funding The study was supported in part by grants from the Ministry of Science and Technology in Taiwan (NSC 102-2410-H-179-014-MY3) to Yu-Kai Chang and (MOST 104-2320-B-010-042) to Hsin-Jen Chen. References 1 World Health Organization . Obesity and Overweight . World Health Organization , 2015 [cited2016 02.13]; http://www.who.int/mediacentre/factsheets/fs311/en/. 2 Ng M , Fleming T , Robinson M et al. . Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013 . Lancet 2014 ; 384 : 766 – 81 . Google Scholar CrossRef Search ADS PubMed 3 Weiss R , Caprio S . The metabolic consequences of childhood obesity . Best Pract Res Clin Endocrinol Metab 2005 ; 19 : 405 – 19 . Google Scholar CrossRef Search ADS PubMed 4 Flynn J . The changing face of pediatric hypertension in the era of the childhood obesity epidemic . Pediatr Nephrol 2013 ; 28 : 1059 – 66 . Google Scholar CrossRef Search ADS PubMed 5 Sonntag D , Ali S , De Bock F . Lifetime indirect cost of childhood overweight and obesity: a decision analytic model . Obesity (Silver Spring, Md) 2016 ; 24 : 200 – 6 . Google Scholar CrossRef Search ADS PubMed 6 Reilly JJ . Descriptive epidemiology and health consequences of childhood obesity . Best Pract Res, Clin Endocrinol Metab 2005 ; 19 : 327 – 41 . Google Scholar CrossRef Search ADS 7 Pan WH , Yeh WT , Chen HJ et al. . The U-shaped relationship between BMI and all-cause mortality contrasts with a progressive increase in medical expenditure: a prospective cohort study . Asia Pac J Clin Nutr 2012 ; 21 : 577 – 87 . Google Scholar PubMed 8 Wabitsch M , Moss A , Kromeyer-Hauschild K . Unexpected plateauing of childhood obesity rates in developed countries . BMC Med 2014 ; 12 : 17 . Google Scholar CrossRef Search ADS PubMed 9 Chung A , Backholer K , Wong E et al. . Trends in child and adolescent obesity prevalence in economically advanced countries according to socioeconomic position: a systematic review . Obes Rev 2016 ; 17 : 276 – 95 . Google Scholar CrossRef Search ADS PubMed 10 Olds T , Maher C , Zumin S et al. . Evidence that the prevalence of childhood overweight is plateauing: data from nine countries . Int J Pediatr Obes 2011 ; 6 : 342 – 60 . Google Scholar CrossRef Search ADS PubMed 11 Ogden CL , Carroll MD , Lawman HG et al. . Trends in obesity prevalence among children and adolescents in the United States, 1988–1994 through 2013–2014 . J Am Med Assoc 2016 ; 315 : 2292 – 9 . Google Scholar CrossRef Search ADS 12 van Jaarsveld CH , Gulliford MC . Childhood obesity trends from primary care electronic health records in England between 1994 and 2013: population-based cohort study . Arch Dis Child 2015 ; 100 : 214 – 9 . Google Scholar CrossRef Search ADS PubMed 13 Visscher TL , Heitmann BL , Rissanen A et al. . A break in the obesity epidemic? Explained by biases or misinterpretation of the data? Int J Obes (Lond) 2015 ; 39 : 189 – 98 . Google Scholar CrossRef Search ADS PubMed 14 Chu NF . Prevalence and trends of obesity among school children in Taiwan—the Taipei Children Heart Study . Int J Obes Relat Metab Disord 2001 ; 25 : 170 – 6 . Google Scholar CrossRef Search ADS PubMed 15 Chen LJ , Fox KR , Haase A et al. . Obesity, fitness and health in Taiwanese children and adolescents . Eur J Clin Nutr 2006 ; 60 : 1367 – 75 . Google Scholar CrossRef Search ADS PubMed 16 Taylor AL , Parento EW , Schmidt L . The increasing weight of regulation: countries combat the global obesity epidemic . Indiana Law J 2015 ; 90 : 257 – 291 . 17 Kersh R . Of nannies and nudges: the current state of U.S. obesity policymaking . Public Health 2015 ; 129 : 1083 – 91 . Google Scholar CrossRef Search ADS PubMed 18 Liou TH , Huang YC , Chou P . Prevalence and secular trends in overweight and obese Taiwanese children and adolescents in 1991–2003 . Ann Hum Biol 2009 ; 36 : 176 – 85 . Google Scholar CrossRef Search ADS PubMed 19 Allman-Farinelli MA , Chey T , Bauman AE et al. . period and birth cohort effects on prevalence of overweight and obesity in Australian adults from 1990 to 2000 . Eur J Clin Nutr 2008 ; 62 : 898 – 907 . Google Scholar CrossRef Search ADS PubMed 20 Tu YK , Chien KL , Burley V et al. . Unravelling the effects of age, period and cohort on metabolic syndrome components in a Taiwanese population using partial least squares regression . BMC Med Res Methodol 2011 ; 11 : 82 . Google Scholar CrossRef Search ADS PubMed 21 Enforcement regulations of physical education for all schools [Database on the Internet]. The Executive Yuan Gazette Online. 2006 [cited 2017/09/15]. http://gazette.nat.gov.tw/egFront/e_detail.do?metaid=9469. 22 Cole TJ , Bellizzi MC , Flegal KM et al. . Establishing a standard definition for child overweight and obesity worldwide: international survey . Br Med J 2000 ; 320 : 1240 – 3 . Google Scholar CrossRef Search ADS 23 Liou YM , Yang YL , Wang TY et al. . School lunch, policy, and environment are determinants for preventing childhood obesity: evidence from a two-year nationwide prospective study . Obes Res Clin Pract 2015 ; 9 : 563 – 72 . Google Scholar CrossRef Search ADS PubMed 24 Wang Y . Cross-national comparison of childhood obesity: the epidemic and the relationship between obesity and socioeconomic status . Int J Epidemiol 2001 ; 30 : 1129 – 36 . Google Scholar CrossRef Search ADS PubMed 25 Wang Y , Monteiro C , Popkin BM . Trends of obesity and underweight in older children and adolescents in the United States, Brazil, China, and Russia . Am J Clin Nutr 2002 ; 75 : 971 – 7 . Google Scholar CrossRef Search ADS PubMed 26 O’Dea JA , Caputi P . Association between socioeconomic status, weight, age and gender, and the body image and weight control practices of 6- to 19-year-old children and adolescents . Health Educ Res 2001 ; 16 : 521 – 32 . Google Scholar CrossRef Search ADS PubMed 27 Jaacks LM , Gordon-Larsen P , Mayer-Davis EJ et al. . Age, period and cohort effects on adult body mass index and overweight from 1991 to 2009 in China: the China Health and Nutrition Survey . Int J Epidemiol 2013 ; 42 : 828 – 37 . Google Scholar CrossRef Search ADS PubMed 28 Andres A , Hull HR , Shankar K et al. . Longitudinal body composition of children born to mothers with normal weight, overweight, and obesity . Obesity (Silver Spring, Md) 2015 ; 23 : 1252 – 8 . Google Scholar CrossRef Search ADS PubMed 29 Oken E , Rifas-Shiman SL , Field AE et al. . Maternal gestational weight gain and offspring weight in adolescence . Obstet Gynecol 2008 ; 112 : 999 – 1006 . Google Scholar CrossRef Search ADS PubMed 30 Wu Q , Suzuki M . Parental obesity and overweight affect the body-fat accumulation in the offspring: the possible effect of a high-fat diet through epigenetic inheritance . Obes Rev 2006 ; 7 : 201 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Nube M . The Asian enigma: predisposition for low adult BMI among people of South Asian descent . Public Health Nutr 2009 ; 12 : 507 – 16 . Google Scholar CrossRef Search ADS PubMed 32 Liu Y , Chen HJ , Liang L et al. . Parent-child resemblance in weight status and its correlates in the United States . PLoS One 2013 ; 8 : e65361 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Public Health Oxford University Press

Age, period and birth cohort effects on the prevalence of overweight and obesity among Taiwanese adolescents: a national population-based study

Loading next page...
 
/lp/ou_press/age-period-and-birth-cohort-effects-on-the-prevalence-of-overweight-ihHgov0RIO
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1741-3842
eISSN
1741-3850
D.O.I.
10.1093/pubmed/fdy025
Publisher site
See Article on Publisher Site

Abstract

Abstract Background The age–period–cohort effects on youth overweight and obesity among junior and senior high school students in Taiwan is not clear. Methods We utilized the 2006–14 nationally representative School Physical Fitness Dataset. Based on the International Obesity Task Force cut-off criteria, a log-linear age–period–cohort analysis was performed to determine the influence of age, period and cohort on the trend in being overweight, obese and severely obese for both sexes. Results The final dataset included 1073173 individuals (n = 520 382 boys and 552 791 girls). For girls, the prevalence of overweight and obesity declined with age, and the prevalence of overweight declined over time. For boys, the prevalence of overweight and obesity declined with age and over time from 2006 to 2014. The prevalence of severe obesity declined over time and increased with age for the boys. The younger birth cohorts had greater odds of being overweight, obese and severely obese than the older birth cohorts. Conclusions After differentiating the age–period–cohort effects, the data suggested a decreasing temporal trend in overweight and obesity among adolescents in Taiwan from 2006 to 2014. Among the birth cohorts of the 1990s, the younger cohorts had greater odds of being overweight and obese than the older cohorts when they reached adolescence. adolescents, age–period–cohort study, body mass index, epidemiology, growth, severe obesity Introduction Overweight and obesity (defined as abnormal or excessive fat accumulation, respectively, with negative impacts on health) has reached epidemic levels in many parts of the world.1,2 Globally, overweight and obesity affected an estimated 23.8% of boys and 22.6% of girls in developed countries in 2013.2 From a public health perspective, tracking the trends in overweight and obesity among youth is important. Excessive weight among youths is associated with medical consequences3,4 and significant financial healthcare costs.5 Furthermore, childhood obesity may increase the susceptibility of children to becoming overweight or obese adults, which consequently leads to more healthcare costs.6,7 Compared with the drastically increasing trend in childhood obesity over the past four decades, emerging data in recent years have suggested that the increase in the prevalence of childhood obesity may have stopped in developed countries.8–12 In the US, the prevalence of childhood obesity remained at ~15–17% from 2005 to 2014.11 UK data showed a similar phenomenon from 2004 to 2013.12 This trend may be related to less obesogenic environments,8 potential method biases, or subgroup heterogeneity.9,13 In Taiwan, some research has examined the prevalence of overweight and obesity in children and adolescents. For instance, 2 and 1% increases among 12–15-year-old boys and girls, respectively, were documented between 1980 and 1994.14 The findings also indicated that the prevalence of overweight and obesity increased from 19.8 to 26.8% and 15.2 to 16.5% among 6–18-year-old boys and girls, respectively, from 1999 to 2001.15 Overall, these statistics imply that the prevalence of overweight and obesity has increased among youths in Taiwan over the past few decades. However, because public attention and policies targeting childhood obesity prevention and control have increased recently,16–18 the prevalence of childhood obesity in Taiwan may plateau, as has been observed in advanced countries.8–12 Importantly, the temporal trend in prevalence may be confounded by varying age distributions and structures of birth cohorts across years. Although there have been reports of temporal trends in the prevalence of overweight and obesity in Taiwan,14,15 these studies are either based on data with a relatively small sample size14 or have not examined the period or birth cohort effects in addition to the age group effects.15 To the best of our knowledge, few studies have discriminated the effects of age, period and cohort on the prevalence of overweight and obesity among the youth population in Taiwan. To untangle the inter-independent problem between these three variables, previous research has utilized the age–period–cohort (APC) analysis model to successfully explore the prevalence of overweight and obesity in adult populations.19,20 Therefore, based on the national School Physical Fitness Database (SPFD) of Taiwan, the aim of the current study was to advance our knowledge on the temporal trend in the prevalence of overweight and obesity in adolescents in Taiwan by examining the influence of age, period and cohort simultaneously. Methods Study population This study was based on the SPFD, which was established by the Taiwan Ministry of Education (MOE) in 2006. The SPFD contains detailed information on the long-term records of physical fitness components of students, such as the body mass index (BMI), cardiovascular fitness, flexibility, and muscular strength and endurance. Physical fitness is measured by physical educational teachers and school nurses who are qualified to perform these measurements following a detailed standardized instruction manual. In addition to fitness information, other demographic data, including the age, sex and grade of each student, are recorded in this dataset. Because uploading fitness data from every student into the SPFD is compulsory according to government laws and regulations,21 the SPFD is representative of the national youth population aged 10–18 years in Taiwan. The original dataset included 1 369 216 height and weight observations. After excluding the observations with missing body weight, height, or birth date values (n = 16 179, 1.18%), unreasonable biological height, weight and/or age values (i.e. age < 60 months, age > 228 months, weight < 10 kg, weight > 200 kg, height < 50 cm or height > 300 cm; n = 3124, 0.23%) and extreme BMI values (>+5 SD or <−5 SD; n = 532, 0.04%), 1 342 282 valid observations remained. Because the majority of the observations were obtained from 2006 to 2014, this study focused on the trends of weight status during this period. Adolescents who were 12–18 years old at the time of the physical examination from 2006 to 2014 were included in the analysis, because this age range contained data collected over a longer period. These children were born between 1994 and 1997. The final sample size was 520 382 boys and 552 791 girls. This study was approved by the local Institutional Ethics Committee at National Taiwan Normal University. Measures The body weights and heights of the children were measured at school by the school nurses, and the BMIs of the adolescents were calculated as the weight (kg)/height2 (m2). The weight statuses were defined using the sex- and age-specific cut-off points from the International Obesity Task Force (IOTF),22 which are recommended for use in international comparisons of the prevalence of overweight and obesity. Based on the IOTF cut-off BMI points, the weight statuses were categorized as follows: overweight, obese and severely obese corresponding to an 18-year-old adult BMI of 25, 30 and 35, respectively. For instance, the international cut-off point for overweight is 21.22 and 21.68 for 12-year-old boys and girls, respectively. In the current study, the obese category encompassed individuals who were obese or severely obese, and the overweight category encompassed individuals who were overweight, obese or severely obese. Statistical analyses The proportion of weight statuses by year of examination and the children’s ages were used to estimate the prevalence of overweight, obesity and severe obesity. Using the Statistical Analysis Software (SAS) GENMOD procedure (SAS 9.3, SAS Institute Inc., Cary, NC, USA) (specifying log link function and Poisson count distribution offset by the denominator population), the Poisson log-linear APC model was applied to analyze binned data of the yearly prevalence. The full model specified all ages, periods (reference = exam years 2009 and 2010) and cohorts (reference = birth year 1995) without an intercept. The deviance difference between the full model and the model omitting the factor of interest was determined to test the effect of each factor on the prevalence by comparing the data with the chi-square distribution to obtain a P-value. The models estimated the prevalence rate ratios for the period (exam year) and cohort (birth year). The prevalence of weight status by age with adjustments for the period and cohort effects was estimated using the least square mean approach. Results Table 1 presents the distributions of overweight, obesity and severe obesity among adolescents aged 12–18 years in Taiwan from 2006 to 2014. The population corresponds to children born from 1994 to 1997. Generally, the prevalence of overweight and obesity increased with the exam year when age was fixed (column-wise trend in Table 1). Additionally, the prevalence of overweight and obesity decreased with age (row-wise trend in Table 1). Table 1 Trends in the prevalence of adolescent overweight, obesity and severe obesity in Taiwan Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Table 1 Trends in the prevalence of adolescent overweight, obesity and severe obesity in Taiwan Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Exam year Girls Boys Age (years) Age (years) 12 13 14 15 16 17 18 12 13 14 15 16 17 18 (A) Sample size  2006 7842 1863 6495 1745  2007 27 534 12 936 2078 23 955 10 298 2115  2008 33 887 30 525 12 843 2112 34 359 27 210 10 411 2165  2009 7962 33 704 27 930 11 735 1933 8622 33 618 24 822 9412 2008  2010 7011 31 668 28 346 13 800 2280 7823 32 692 25 445 10 708 2114  2011 7736 36 183 29 584 12 344 1866 7961 36 562 25 687 10 323 2045  2012 6218 33 706 30 117 13 427 6846 35 251 26 912 10 986  2013 5874 33 226 29 597 6316 34 556 26 442  2014 4778 10 146 5030 9448 (B) Prevalence of overweight  2006 22.1% 20.8% 32.1% 33.8%  2007 23.2% 20.6% 15.7% 34.0% 29.5% 30.3%  2008 21.9% 20.0% 17.3% 13.8% 36.5% 31.0% 26.2% 26.9%  2009 23.2% 19.8% 17.5% 14.9% 14.1% 37.1% 34.0% 28.4% 24.1% 23.9%  2010 21.2% 17.9% 15.7% 13.7% 11.3% 36.1% 32.5% 26.7% 23.5% 22.5%  2011 19.6% 16.2% 13.5% 12.2% 10.6% 33.6% 29.8% 24.6% 20.4% 20.5%  2012 16.5% 14.1% 12.4% 11.9% 30.4% 27.3% 21.7% 18.1%  2013 14.5% 13.4% 12.7% 27.1% 25.3% 21.3%  2014 14.3% 12.9% 25.1% 22.9% (C) Prevalence of obesity  2006 4.0% 3.5% 7.8% 9.0%  2007 4.7% 3.8% 2.1% 8.5% 6.4% 7.9%  2008 4.4% 4.2% 3.4% 2.5% 10.5% 7.3% 5.6% 6.7%  2009 5.1% 4.1% 3.5% 3.3% 2.8% 10.3% 9.5% 7.5% 5.4% 6.1%  2010 4.5% 3.7% 3.4% 2.8% 1.9% 11.2% 9.7% 7.0% 5.4% 5.7%  2011 4.4% 3.7% 3.1% 2.9% 2.0% 10.6% 9.2% 6.5% 4.4% 5.8%  2012 3.5% 3.3% 3.0% 2.9% 9.8% 8.3% 5.6% 4.0%  2013 3.4% 3.2% 3.4% 8.3% 7.7% 5.7%  2014 3.5% 3.3% 7.8% 7.1% (D) Prevalence of severe obesity  2006 0.4% 0.5% 0.8% 0.6%  2007 0.6% 0.6% 0.4% 1.0% 0.7% 1.1%  2008 0.6% 0.6% 0.5% 0.3% 1.3% 0.7% 0.7% 0.9%  2009 0.6% 0.5% 0.5% 0.6% 0.4% 1.5% 1.2% 1.0% 0.7% 1.3%  2010 0.5% 0.5% 0.6% 0.6% 0.4% 1.5% 1.6% 1.1% 0.8% 1.0%  2011 0.6% 0.7% 0.6% 0.5% 0.4% 1.6% 1.7% 1.2% 0.9% 1.1%  2012 0.6% 0.6% 0.6% 0.6% 1.9% 1.7% 1.1% 0.9%  2013 0.8% 0.7% 0.7% 1.9% 1.9% 1.3%  2014 0.9% 0.8% 1.6% 1.8% Table 2 presents whether age, period (exam years) and/or cohort (birth years) affect the prevalence of overweight, obesity and severe obesity. The global tests showed the significance of the age, period and cohort effects. The direction of the temporal effect was revealed by delineation of the prevalence by age, time and cohort when the other two temporal factors were controlled, as shown in Figs 1 and 2. The findings are described by the age, period and cohort effects as follows. Table 2 Testing the age, period and cohort effects for the three weight statuses stratified by sex Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Note: The individual effect of age, period or cohort was assessed by comparing the difference in deviance between the full three-factor model (age–period–cohort) and the two-factor model that omitted the factor of interest. Table 2 Testing the age, period and cohort effects for the three weight statuses stratified by sex Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Model df Deviance Effect Difference df Difference deviance P value Girls  Overweight   Age–period–cohort 10 9.6   Age–period 13 24.4 Cohort 3 14.9 0.002   Age–cohort 17 48.0 Period 7 38.5 <0.001   Period–cohort 16 117.2 Age 6 107.6 <0.001  Obesity   Age–period–cohort 10 9.1   Age–period 13 18.0 Cohort 3 8.9 0.031   Age–cohort 17 16.2 Period 7 7.1 0.422   Period–cohort 16 48.3 Age 6 39.2 <0.001  Severe obesity   Age–period–cohort 10 7.6   Age–period 13 9.9 Cohort 3 2.2 0.525   Age–cohort 17 20.4 Period 7 12.8 0.077   Period–cohort 16 17.2 Age 6 9.5 0.146 Boys  Overweight   Age–period–cohort 10 13.4   Age–period 13 174.0 Cohort 3 160.6 <0.001   Age–cohort 17 47.4 Period 7 34.0 <0.001   Period–cohort 16 50.9 Age 6 37.5 <0.001  Obesity   Age–period–cohort 10 21.7   Age–period 13 271.8 Cohort 3 250.1 <0.001   Age–cohort 17 56.6 Period 7 34.9 <0.001   Period–cohort 16 36.8 Age 6 15.1 0.020  Severe obesity   Age–period–cohort 10 6.2   Age–period 13 101.1 Cohort 3 94.8 <0.001   Age–cohort 17 27.6 Period 7 21.4 0.003   Period–cohort 16 27.3 Age 6 21.1 0.002 Note: The individual effect of age, period or cohort was assessed by comparing the difference in deviance between the full three-factor model (age–period–cohort) and the two-factor model that omitted the factor of interest. Fig. 1 View largeDownload slide Estimated age, period and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent girls: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends show risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 1 View largeDownload slide Estimated age, period and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent girls: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends show risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 2 View largeDownload slide Estimated age, period, and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent boys: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends are risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Fig. 2 View largeDownload slide Estimated age, period, and cohort trends for the prevalence of overweight (a), obesity (b) and severe obesity (c) among adolescent boys: Taiwan 2006–14. Age trends are the estimated prevalence by age and year. Period and cohort trends are risk ratios compared to the reference exam year (2010) and reference birth year (1995), respectively. Age effect: The age effect significantly affected the prevalence of overweight and obesity for both girls and boys. According to Table 1 and Figs 1 and 2, the prevalence of overweight and obesity decreased with age in the adolescent population (both sexes). The age trend for severe obesity was significant only for boys. Period effect: Regarding the period effect, the prevalence of overweight, obesity and severe obesity for boys decreased over time from 2006 to 2014. For girls, only the prevalence of overweight varied by exam year. The prevalence of overweight in girls was increased by 1.12-fold (95% confidence interval [CI]: 1.01–1.24) and 1.13-fold (95% CI: 1.05–1.20) in 2006 and 2007, respectively, when compared with the prevalence of overweight from 2009 to 2010. For boys, the prevalence rates of overweight, obesity, and severe obesity all varied significantly by exam year (Table 2). Compared with the prevalence of overweight from 2009 to 2010, the prevalence of overweight in 2006 was increased by 1.18-fold (95% CI: 1.08–1.28), and the prevalence of overweight in 2014 was reduced by 0.88-fold (95% CI: 0.80–0.97). Compared with the prevalence of obesity in boys from 2009 to 2010, the prevalence of obesity in 2006 was increased by 1.45-fold (95% CI: 1.23–1.71), and the prevalence in 2014 was reduced by 0.81-fold (95% CI: 0.67–0.98). Similarly, compared with the prevalence of severe obesity in boys from 2009 to 2010, the prevalence of severe obesity in 2006 was increased by 1.69-fold (95% CI: 1.07–2.66), and the prevalence in 2014 was reduced by 0.62-fold (95% CI: 0.38–1.01). Generally, the results suggested a decreasing prevalence of overweight, obesity and severe obesity among boys. Cohort effect: The prevalence of overweight, obesity, and severe obesity varied significantly by birth year. The prevalence rates of overweight and obesity were greater for children born in the later years than those for children born in the earlier years (Figs 1 and 2). Compared with boys born in 1994, boys born in 1997 had 1.35 (95% CI: 1.27–1.44) times, 2.01 (95% CI: 1.78–2.28) times and 2.91 (95% CI: 2.09–4.05) times higher prevalence rates of overweight, obesity, and severe obesity, respectively. Compared with girls born in 1994, girls born in 1997 had 1.15 (95% CI: 1.07–1.25) times and 1.25 (95% CI: 1.05–1.48) times higher prevalence rates of overweight and obesity, respectively. Discussion Main findings of this study This study examined eight years of national data from the SPFD (2006–14) for children and adolescents in Taiwan and provided unique comparative information regarding trends in the prevalence of preadolescent and adolescent overweight, obesity and severe obesity (Tables 1 and 2). The prevalence of overweight and obesity among adolescents has stabilized or even decreased during the last decade. The prevalence of overweight and obesity decreased with age among Taiwanese adolescents, but the prevalence of severe obesity increased with age. Interestingly, the younger birth cohorts were more likely to be overweight or obese than the older birth cohorts. What is already known on this topic Regardless of the definition used for overweight or obesity, data from 1990 to 2000s indicated that the prevalence was increasing in Taiwan. In Taipei (the capital city of Taiwan), the prevalence of obesity among 12–15-year-old school boys significantly increased from 12.4 to 16.4% from 1980 to 1994, respectively.14 A similar increasing trend in the prevalence of overweight and obesity was reported among students aged 6–18 years from 1999 to 2001.15 Based on the representative data from all Taiwanese adolescents (n = 38 921), the prevalence of overweight increased from 14.1 to 18.6% and 12.8 to 13.0% in boys and girls, respectively.15 The prevalence of obesity also increased from 5.7 to 8.2% and 2.4 to 3.6% in boys and girls, respectively.15 What this study adds Similar to data from economically advanced countries or areas,8–10 such as the USA11 and UK,12 the prevalence of childhood obesity in Taiwan appeared to stop increasing. From 2006 to 2014, the prevalence of overweight and obesity decreased among adolescents aged 12–18 years. Although the exact explanation for this recent decline is unclear, policy measures to prevent and control the obesity epidemic have been in place since the mid-2000s. In 2005, the Ministry of Education issued an executive order that banned the sale of sweetened beverages and unhealthy snacks in elementary and junior high schools. Additionally, the Health-Promoting School Program, which includes promotion of a healthy weight status by improving school lunch quality, offering nutrition education, and encouraging children’s healthy lifestyle (e.g. sleep, balance diet and active life), was completely implemented in all elementary and junior high schools in Taiwan by 2008.23 In 2013, the Taiwanese Food and Drug Administration has implemented the regulations regarding to the advertisement of foods that ‘are unsuitable for long-term consumption for children’ to children.17 These policies could have created an environment that counteracted the rising trend in overweight and obesity in adolescents. Further research is warranted to elucidate the impacts of these policies. The age trend in the prevalence of overweight and obesity can reveal the age trajectory of the weight status during life stages. Our current results indicate inverse relationships between overweight and obese statuses and age. The finding agrees with earlier studies from Taiwan14,15,18,24 and China24,25 that found that younger children were at greater risk of being overweight and obese than older children. For example, the prevalence of overweight/obesity among boys declined from 27.6 to 20.8%, and a similar decrease in the prevalence of overweight/obesity from 17.1 to 8.7% was observed among girls.15 The reduction in the prevalence with increased age might reflect the increased weight-related concerns or body dissatisfaction among older adolescents,26 which might subsequently lead to an increase in the practice of weight-loss behaviors.26 However, the decreased prevalence of overweight and obesity with age in Taiwan and China also raises a question concerning the application of the international growth references for these Asian populations. For instance, the prevalence of overweight and obesity that appears to be decreasing with age could result from slowing of the weights and heights of the target populations during adolescence when compared with the reference population. Although the IOTF reference incorporates data from Hong Kong, further research is needed to examine the fitness of international growth references to these Asian populations. Another finding that deserves attention is that the prevalence of severe obesity increases with age. This finding implies that there could be different groups of adolescents with diverse growth trajectories during this life stage. Although the majority of adolescents had lower risks of overweight and obesity while growing up, a small proportion might need additional medical care to prevent the severe obesity status from carrying over to adulthood. Relative to the birth cohort born during 1994, the cohort born toward the end of 1997 tended to experience increased odds of being overweight and obese. Another study on Chinese adults also showed that the younger cohort exhibited significantly increased BMI values compared to the older cohort.27 Birth cohort effects reflect environmental changes, such as maternal increased food consumption and sedentary behaviors, during the economic transition. The per capita gross domestic product increased from $8 132 in 1990 to $14 519 in 2000. This economic growth in Taiwan over recent decades might have resulted in an ‘obesogenic’ environment, with a lower cost of energy-dense foods and more convenient lifestyle,19 since the prevalence of obesity was rising during this period as mentioned above. Maternal obesity and diet quality might affect the offspring’s future risks of obesity.28–30 The modified risks of obesity due to the intrauterine environment might explain the observed cohort differences in obesity. The younger birth cohorts’ greater prevalence of overweight and obesity might reflect their exposure to the increasingly obesity-promoting environment when they were fetuses during the 1990s. Limitations of this study Some limitations of the current study should be considered. First, we were unable to examine township and city district-specific trends in overweight, obesity and severe obesity. Prior research has suggested that the prevalence of obesity is more severe in poor and rural areas than that in the more urbanized areas of Taiwan.23 Additionally, interventional strategies against overweight and obesity might be developed and implemented at the regional level. Understanding which of these policies are most effective is not possible by examining the national data used in this study. Including geographic information in future research will increase our understanding of the geographic prevalence of overweight and obesity in Taiwan and the effectiveness of interventional strategies. Another limitation was the lack of comprehensive information for the students in the current study (related to family and ethnic groups). Some research has suggested that the prevalence may be influenced by ethnic groups.31,32 Because 92.7% of Taiwanese are of the Han nationality, the current findings provide a general picture of the effects of the age, period and birth cohort on the prevalence of overweight and obesity in Taiwan. Third, the criteria for overweight, obesity and severe obesity in the current study were defined based on the BMI values, which is a popular and straightforward measurement method designed for epidemic research. However, the use of an adiposity measure, such as skinfold thickness, should be considered in future studies. Fourth, the data were originally intended for administrative purposes, and the measurements were taken by school personnel. Despite the standardized protocol, measurement errors could still have occurred. However, the measurement errors should be random and canceled out when the data for the whole population are pooled. Moreover, there was no evidence of population-level systematic errors in the body weight and height measurements. Fifth, the ages of the children in this study spanned from junior to senior high school. Compulsory education in Taiwan extends for 9 years until junior high school. Thus, not all junior high school students go on to senior high school. Nevertheless, the enrollment rates were high for the junior and senior high schools. For example, in 2008, 97% of children aged 12–14 years were enrolled in junior high school, and 91% of children aged 15–17 years were enrolled in senior high school. In 2014, 98% of 12–14-year-old children were enrolled in junior high school, and 94% of 15–17-year-old children were enrolled in senior high school. The small differences in the enrollment rates between junior and senior high school cannot truly explain the age trend in overweight and obesity. Despite these limitations, our current study possesses several strengths, such as the large nationally representative dataset, which is free from self-report bias through standardized collection protocols for the anthropometric measurements of height and weight. More importantly, the current study utilized a refined statistical approach (the APC analysis) to overcome the inherent problem of discriminating the effects of age, period and birth cohort and enabled us to demonstrate the independent effects of age, period and birth cohort on the prevalence of overweight and obesity among Taiwanese youths. In conclusion, to the best of our knowledge, the prevalence of overweight and obesity among Taiwanese youth has stopped increasing and even started declining. Additionally, the prevalence of overweight and obesity decreased significantly as the children aged. Nevertheless, the prevalence of severe obesity increased with age. Apart from the contribution of age, clear cohort differences were present for overweight and obesity among Taiwanese youths. As noted, the cohort differences may reflect differences in the years of exposure to increasing environmental obesogenecity during the 1990s among these birth cohorts. Although the prevalence in Taiwan did not increase from 2006 to 2014, continual monitoring of the obesogenic environment and weight status of children is necessary to understand the influence of policies on temporal trends. Conflicts of interest The authors declared no conflicts of interests. Acknowledgements We thank Drs Chung-Ching Chen and Tzyy-Yuang Shiang for preparing the article. Authors’ contributions YKC and HJC conceived the study and designed the analysis processes. CHC and HJC were responsible for the data cleaning, analysis and interpretation. YKC, HJC and CHC drafted sections of the article. All authors were involved in manuscript review and revision and approved the final version of this article. Funding The study was supported in part by grants from the Ministry of Science and Technology in Taiwan (NSC 102-2410-H-179-014-MY3) to Yu-Kai Chang and (MOST 104-2320-B-010-042) to Hsin-Jen Chen. References 1 World Health Organization . Obesity and Overweight . World Health Organization , 2015 [cited2016 02.13]; http://www.who.int/mediacentre/factsheets/fs311/en/. 2 Ng M , Fleming T , Robinson M et al. . Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013 . Lancet 2014 ; 384 : 766 – 81 . Google Scholar CrossRef Search ADS PubMed 3 Weiss R , Caprio S . The metabolic consequences of childhood obesity . Best Pract Res Clin Endocrinol Metab 2005 ; 19 : 405 – 19 . Google Scholar CrossRef Search ADS PubMed 4 Flynn J . The changing face of pediatric hypertension in the era of the childhood obesity epidemic . Pediatr Nephrol 2013 ; 28 : 1059 – 66 . Google Scholar CrossRef Search ADS PubMed 5 Sonntag D , Ali S , De Bock F . Lifetime indirect cost of childhood overweight and obesity: a decision analytic model . Obesity (Silver Spring, Md) 2016 ; 24 : 200 – 6 . Google Scholar CrossRef Search ADS PubMed 6 Reilly JJ . Descriptive epidemiology and health consequences of childhood obesity . Best Pract Res, Clin Endocrinol Metab 2005 ; 19 : 327 – 41 . Google Scholar CrossRef Search ADS 7 Pan WH , Yeh WT , Chen HJ et al. . The U-shaped relationship between BMI and all-cause mortality contrasts with a progressive increase in medical expenditure: a prospective cohort study . Asia Pac J Clin Nutr 2012 ; 21 : 577 – 87 . Google Scholar PubMed 8 Wabitsch M , Moss A , Kromeyer-Hauschild K . Unexpected plateauing of childhood obesity rates in developed countries . BMC Med 2014 ; 12 : 17 . Google Scholar CrossRef Search ADS PubMed 9 Chung A , Backholer K , Wong E et al. . Trends in child and adolescent obesity prevalence in economically advanced countries according to socioeconomic position: a systematic review . Obes Rev 2016 ; 17 : 276 – 95 . Google Scholar CrossRef Search ADS PubMed 10 Olds T , Maher C , Zumin S et al. . Evidence that the prevalence of childhood overweight is plateauing: data from nine countries . Int J Pediatr Obes 2011 ; 6 : 342 – 60 . Google Scholar CrossRef Search ADS PubMed 11 Ogden CL , Carroll MD , Lawman HG et al. . Trends in obesity prevalence among children and adolescents in the United States, 1988–1994 through 2013–2014 . J Am Med Assoc 2016 ; 315 : 2292 – 9 . Google Scholar CrossRef Search ADS 12 van Jaarsveld CH , Gulliford MC . Childhood obesity trends from primary care electronic health records in England between 1994 and 2013: population-based cohort study . Arch Dis Child 2015 ; 100 : 214 – 9 . Google Scholar CrossRef Search ADS PubMed 13 Visscher TL , Heitmann BL , Rissanen A et al. . A break in the obesity epidemic? Explained by biases or misinterpretation of the data? Int J Obes (Lond) 2015 ; 39 : 189 – 98 . Google Scholar CrossRef Search ADS PubMed 14 Chu NF . Prevalence and trends of obesity among school children in Taiwan—the Taipei Children Heart Study . Int J Obes Relat Metab Disord 2001 ; 25 : 170 – 6 . Google Scholar CrossRef Search ADS PubMed 15 Chen LJ , Fox KR , Haase A et al. . Obesity, fitness and health in Taiwanese children and adolescents . Eur J Clin Nutr 2006 ; 60 : 1367 – 75 . Google Scholar CrossRef Search ADS PubMed 16 Taylor AL , Parento EW , Schmidt L . The increasing weight of regulation: countries combat the global obesity epidemic . Indiana Law J 2015 ; 90 : 257 – 291 . 17 Kersh R . Of nannies and nudges: the current state of U.S. obesity policymaking . Public Health 2015 ; 129 : 1083 – 91 . Google Scholar CrossRef Search ADS PubMed 18 Liou TH , Huang YC , Chou P . Prevalence and secular trends in overweight and obese Taiwanese children and adolescents in 1991–2003 . Ann Hum Biol 2009 ; 36 : 176 – 85 . Google Scholar CrossRef Search ADS PubMed 19 Allman-Farinelli MA , Chey T , Bauman AE et al. . period and birth cohort effects on prevalence of overweight and obesity in Australian adults from 1990 to 2000 . Eur J Clin Nutr 2008 ; 62 : 898 – 907 . Google Scholar CrossRef Search ADS PubMed 20 Tu YK , Chien KL , Burley V et al. . Unravelling the effects of age, period and cohort on metabolic syndrome components in a Taiwanese population using partial least squares regression . BMC Med Res Methodol 2011 ; 11 : 82 . Google Scholar CrossRef Search ADS PubMed 21 Enforcement regulations of physical education for all schools [Database on the Internet]. The Executive Yuan Gazette Online. 2006 [cited 2017/09/15]. http://gazette.nat.gov.tw/egFront/e_detail.do?metaid=9469. 22 Cole TJ , Bellizzi MC , Flegal KM et al. . Establishing a standard definition for child overweight and obesity worldwide: international survey . Br Med J 2000 ; 320 : 1240 – 3 . Google Scholar CrossRef Search ADS 23 Liou YM , Yang YL , Wang TY et al. . School lunch, policy, and environment are determinants for preventing childhood obesity: evidence from a two-year nationwide prospective study . Obes Res Clin Pract 2015 ; 9 : 563 – 72 . Google Scholar CrossRef Search ADS PubMed 24 Wang Y . Cross-national comparison of childhood obesity: the epidemic and the relationship between obesity and socioeconomic status . Int J Epidemiol 2001 ; 30 : 1129 – 36 . Google Scholar CrossRef Search ADS PubMed 25 Wang Y , Monteiro C , Popkin BM . Trends of obesity and underweight in older children and adolescents in the United States, Brazil, China, and Russia . Am J Clin Nutr 2002 ; 75 : 971 – 7 . Google Scholar CrossRef Search ADS PubMed 26 O’Dea JA , Caputi P . Association between socioeconomic status, weight, age and gender, and the body image and weight control practices of 6- to 19-year-old children and adolescents . Health Educ Res 2001 ; 16 : 521 – 32 . Google Scholar CrossRef Search ADS PubMed 27 Jaacks LM , Gordon-Larsen P , Mayer-Davis EJ et al. . Age, period and cohort effects on adult body mass index and overweight from 1991 to 2009 in China: the China Health and Nutrition Survey . Int J Epidemiol 2013 ; 42 : 828 – 37 . Google Scholar CrossRef Search ADS PubMed 28 Andres A , Hull HR , Shankar K et al. . Longitudinal body composition of children born to mothers with normal weight, overweight, and obesity . Obesity (Silver Spring, Md) 2015 ; 23 : 1252 – 8 . Google Scholar CrossRef Search ADS PubMed 29 Oken E , Rifas-Shiman SL , Field AE et al. . Maternal gestational weight gain and offspring weight in adolescence . Obstet Gynecol 2008 ; 112 : 999 – 1006 . Google Scholar CrossRef Search ADS PubMed 30 Wu Q , Suzuki M . Parental obesity and overweight affect the body-fat accumulation in the offspring: the possible effect of a high-fat diet through epigenetic inheritance . Obes Rev 2006 ; 7 : 201 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Nube M . The Asian enigma: predisposition for low adult BMI among people of South Asian descent . Public Health Nutr 2009 ; 12 : 507 – 16 . Google Scholar CrossRef Search ADS PubMed 32 Liu Y , Chen HJ , Liang L et al. . Parent-child resemblance in weight status and its correlates in the United States . PLoS One 2013 ; 8 : e65361 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Journal of Public HealthOxford University Press

Published: Feb 23, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off