September 12, 2013|Categories: Analytics and Research
Iyue Sung, Director of Core Analytics, and Rick McKellar, an intern on the Core Analytics Team, also contributed to this report.
When Washington Post writer Lydia DePillis recently released a blog post entitled “Looks like the U.S. is winning its war on childhood obesity,” the athenaResearch team read her work with great interest. In the post, DePillis describes encouraging reductions in obesity, in select U.S. counties from 2007 to 2009, and cites work by the Robert Wood Johnson Foundation showing further improvements in Boston, Philadelphia, New York, and a number of other cities since around 2006. The
Centers for Disease Control (CDC) similarly cited a reduction in obesity in low income children aged two to four for the time period 2008 - 2011.
Our Cloud Analytics Team decided to explore how childhood obesity trends are playing out across our own national network, currently comprising clinical records for approximately 15 million patients. Because we use a cloud-based, continually updated platform, we can measure childhood obesity based on data coming from pediatricians on a daily basis, and report findings through the end of June 2013.
Two caveats up front: First, we measure obesity differently than most surveys and clinical studies, which typically either ask parents for self-reported height and weight through phone surveys, or simply take biometric measurements at schools or in other settings. In contrast, we report on children’s height and weight readings obtained during their encounters with providers on the athenahealth national network. Specifically, the data is gathered on an ongoing basis directly from athenaClinicals, our cloud-based electronic health record (EHR).
We follow the CDC’s definition of childhood obesity; if any child has a body mass index (BMI) above the CDC threshold at any point during the year, we consider him or her obese.
Second caveat: Our sample is drawn from athenahealth practices around the country, but may not be as large as samples used in other studies.
On to the data. The chart below is based on BMI measures recorded during 617,000 pediatric visits of middle school students (age 12-14) and high school students (age 15-18) from 490 ambulatory medical practices between July 2010 and June 2013. (Each practice used athenahealth’s EHR for all three years studied. Including all practices with pediatric visits during these three years raised the visit count to approximately one million and showed similar trends.)
We observed that some geographies experienced declines in adolescent obesity. Unfortunately, we also found that the overall trend in obesity within the athenahealth network for middle school- and high school-age patients has remained essentially flat for the past three years.
For middle schoolers, the obesity rate hovered between 21.9% and 22.3%; for high school students, between 19.5% and 19.8%. The lower obesity rate for the older group is at least in part attributable to the CDC’s definition of childhood obesity, which varies by age.
In short, local declines in childhood obesity within the athenahealth network are occurring against a national backdrop where obesity rates are flat. If there is indeed a trend towards declining childhood obesity rates nationally, we are not seeing it at athenahealth practices.
For those interested in a deeper analysis, the underlying data are summarized below:
This data underscores the importance of the work being done on a local level to control childhood obesity. The local success stories are proof that obesity can be managed; these efforts need to be scaled to the national level.
We will continue to monitor childhood obesity and post again the minute we see changes. In addition, we look forward to providing analytic support for BOKS, an initiative of Reebok and the Reebok Foundation, to expand free before-school physical activity programs for children.
Please send along any questions, comments or suggestions for follow-up analyses to firstname.lastname@example.org, submit a comment or join the discussion on Twitter by following #athenadata. We would be especially interested to hear from obesity researchers.