Statistical Issues of Measuring and Monitoring Childhood Obesity from Administrative Records: The New York City FITNESSGRAM

• Activity Number: 34
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Kevin J. Konty*+
• Companies: New York City Department of Health and Mental Hygiene
• Keywords: model-based estimation ; administrative records ; health monitoring ; obesity ; public health practice ; missing data

Abstract

Childhood obesity is the focus of a variety of public health initiatives. One basic challenge in such efforts is the measurement and monitoring of obesity. In NYC, a fitness assessment implemented in public schools includes height and weight measures linkable across years and has the potential to provide both cross-sectional and longitudinal estimates.

We use model-based estimation that predicts obesity for all students in NYC public schools using information from the enrollment database. Although common in statistics, our approach is rare in public health practice. Statistical issues such as population coverage, variance estimation, identification of trends, and sensitivity of estimates to poor data quality are common in many uses of administrative datasets for health monitoring, including sentinel systems and electronic health records. Extensions of our approach will be discussed.