Tuesday, August 9, 2011

Colorado Childhood Obesity Data by Region 2010

In response to recent media requests, Renee Calanan, EPE,generated a spreadsheet of updated Colorado Childhood Obesity Data by Region, Race/Ethnicity (state level), and Year (state level). These data represent the most recent data available.

In the spreadsheet, Renee included some data interpretation within the 3 tabs of the spreadsheet. In general, Renee notes that "it is difficult to make any interesting conclusions about the data by region of Colorado. If you look at the prevalence estimates only, there is variation by region… but it is important to also check the confidence intervals." 

General Overview
Regarding the trend data, there was not a significant change in childhood obesity among 2-14 year olds from 2004 to 2010. Although there was not a significant increase, there also was not a significant decrease, which is what we would like to see. I cannot explain why there was a significant increase seen for Colorado in other reports (i.e., F as in Fat; your Health Policy Solutions article from February 9, 2011) – except that there are differences between the data I sent you and those reports: the data I sent you includes children aged 2-14 years and uses Colorado Child Health Survey (CHS) data; the other reports include children aged 10-17 years and use a different data source, National Survey of Children’s Health (NSCH) data. CHS data is not directly
comparable to NSCH data or any other national data.

For further information, take a look at CDPHE’s most recent report on obesity (published in 2009) located at http://www.cdphe.state.co.us/pp/COPAN/ObesityReport.pdf

Notes on interpreting the spreadsheet data
The prevalence estimates with confidence intervals highlighted in orange should be interpreted with caution because the confidence intervals are so large. This is due to small sample sizes (ultimately because of the small population within those regions). Also, please note that a conservative statistical test to determine whether two prevalence estimates are significantly different from one another is to compare their confidence intervals. If the two confidence intervals do not overlap, then they are significantly different from one another.

No comments:

Post a Comment