Discussion
Severe childhood obesity is a subset of childhood obesity that is associated with serious short- and long-term health consequences and is increasing in prevalence among US children. With its potentially grave consequences, it is important that the subgroup of severely obese children be monitored to inform efforts to reduce the prevalence and prevent incidence of severe obesity. Although it is known that childhood obesity has decreased in NYC, the prevalence and trends of severe childhood obesity has not been examined. Ours is the first study to do so.
NYC public elementary and middle school students experienced a decrease of 9.5% in severe obesity from 2006–07 to 2010–11. Among this same population, obesity also decreased during this period by 5.5%, suggesting that the public response to the obesity epidemic is affecting all levels of childhood obesity. We found many similarities in the patterns of prevalence of severe obesity and obesity; both were highest among minority, poor, and male children. However, the prevalence of severe obesity increased with age while the prevalence of obesity peaked in those aged 7 to 10 years and then declined. As age increased, higher proportions of obese students became severely obese. Similar to obesity, severe obesity declined the most among the lowest risk populations (nonminority and wealthy students). We also found that including high BIVs could have a significant effect on prevalence estimates of severe obesity. Our work establishes a baseline for future monitoring of the severely obese pediatric population.
Severe obesity in children is not routinely monitored, despite its conferring the highest risk for poor health outcomes. It is important to monitor severe obesity independently of obesity to accurately identify potentially differing trends between the 2 groups. Although we did not find many differing patterns in the prevalence and trends between severe obesity and obesity, the differences by age indicate the potential for other differences that should continue to be monitored. Also, decreases in severe obesity represent an important public health outcome independent of decreases in overall prevalence of obesity. Such decreases may go undetected if we monitor obesity alone, because the severely obese who become obese will continue to be classified as obese. Although the overall goal is to reduce the rate of obesity, a reduction in the rate of severe obesity, even in the context of an increase in overall obesity, has important implications for the health of high-risk populations.
Our data are similar to results from a study of Philadelphia public school children that found obesity rates of 20.5% and severe obesity rates of 7.9%. That study also found that groups with the highest rates of severe obesity and obesity at the beginning of the study had the smallest decreases over time. A southern California cohort reported a prevalence of 6.4% among children aged 2 to 19 years enrolled in a prepaid health plan from 2007 through 2008 whose demographic distributions were similar to the southern California census. Another study classified children having a BMI at or greater than 120% of the 95th percentile as "Class 2 Obesity" by using National Health and Nutrition Examination Survey data from 1999–2012 for children aged 2 to 19 years. This study reported similar cross-sectional patterns by age, sex, and race/ethnicity with the highest proportions in prevalence found among older (aged 12–19 y), male, and minority children; however, severe obesity prevalence increased from 3.8% in the 1999 through 2000 study period to 5.9% in the 2011 through 2012 period.
The CDC's 2000 growth charts use a standardized distribution derived from populations of children measured from the 1970s through 1990s when 5% of children were declared obese. These growth charts were not designed to display percentiles beyond 97%. The flags for BIV were meant to flag data entry or measurement errors. However, because childhood obesity increased from 5% to nearly 20% from 1980 through 2012, many plausible values are likely for BMI measures that are above the z score upper limit defined in the 2000 CDC growth charts. In our analysis, BIVs accounted for 2% of all measurements in the 2010–11 school year. If all high BIVs were found to be plausible, the prevalence of severe obesity among NYC children in grades K–8 would increase by as much as 15%, and the prevalence of obesity would increase by 4%.
Our study has several limitations. The first is that our results do not include private and charter schools, which constitute around 30% of NYC children; these children are not required to participate in NYC FITNESSGRAM. Another limitation is that NYC FITNESSGRAM coverage was low in early years (60.9% in the 2006–07 school year); however, coverage increased each year and was 93.0% in 2010–11. We address this by weighting the measured population to be representative of the enrolled population accounting for individual- and school-level characteristics. Although we addressed potential issues with BIV exclusion criteria, which is biased toward lower severe obesity prevalence, it is beyond the scope of this analysis to test new methods of identifying outliers among childhood BMI data. Lastly, the data presented here are cross-sectional. Analyses that examine the BMI trajectories of children with longitudinal records would be able to more accurately determine the true trend of obesity and severe obesity as well as movement into BIVs.
Our study has several strengths. First, the study population was large and diverse (contributing over 2 million BMI measurements), which allowed us to examine severe obesity and obesity by subgroup. Additionally, we examined 5 years of data, which allowed us to examine trends over time. Finally, we were able to show the potential effect of the current BIV definition and to demonstrate the importance of a revised definition that obtains wide acceptance among childhood obesity researchers.
Although obesity is beginning to plateau among US children, severe childhood obesity overall is increasing nationally, but not in NYC, as our study showed, where severe obesity is decreasing along with or faster than obesity. Accurately monitoring severely obese populations provides clinicians and public health officials with the ability to target those most at risk for poor health outcomes. Furthermore, given the increased morbidity risks for the severely obese, preventing progression from obesity to severe obesity becomes all the more important. Policymakers should make an effort to follow the trends in both severe obesity and obesity while combating the overall problem of childhood obesity. Efforts to decrease overall childhood obesity prevalence in NYC have been successful at decreasing the subpopulation of severely obese children as well.