[The effects of age, period, and birth cohorts on the rates of overweight and underweight students in 1977-2006 and a prediction of the rates in 2007-2016].

OBJECTIVES: The increase in the overweight or underweight student population in Japanese schools is an important health issue. To assess the independent effects of age, period, and birth cohorts, we analyzed trends in the rates of overweight and underweight students from 1977 to 2006 using a Bayesian age-period-cohort (APC) analysis. Additionally, we predicted the rates overweight and underweight students in 2007-2016. METHODS: We created a data set of the rates of overweight and underweight students aged 6-14 years using the annual school health survey report data. We then analyzed a cohort table that plotted age against calendar time using a Bayesian APC model. We also made a prediction of the rates of overweight and underweight students in 2007-2016. RESULTS: For overweight students, the age effect increased from 6 to 11 years of age for male students and from 6 to 12 years of age for female students; thereafter, the effects decreased. The period effects consistently increased until late 1990, and decreased thereafter for both male and female students. The cohort effects increased for male students born between 1963 and 1969, and later decreased for those born in 1981. However, this trend for male students later increased. For female students, the cohort effects decreased for those born between 1963 and 1975, and later increased for those born in 1990, indicating a plateau or slight increase in the trend. For underweight students, the age effect in male students increased from 7 to 10 years of age, and then plateaued; whereas, the age effect increased from 7 to 12 years of age for female students before reaching a plateau. The period effects increased consistently by 2000 and decreased slightly in both male and female students. The birth cohort effect in male students increased for those born after the mid-1980s, and then plateaued in the early 1990s. The cohort effect increased for female students born after 1984, and then plateaued in 1993. The projections for the rates of overweight and underweight student population indicated a steady trend until 2016. CONCLUSION: The rate of overweight and underweight students was strongly influenced by age; however, period and birth cohort also played a role. The projections for the rates of overweight and underweight students indicated a steady trend until 2016. These results suggest that strategies based on age-, period-, and cohort-specific measures may be required for future interventions for preventing overweight and underweight among students.