Using local data to monitor obesity rates in Wisconsin counties, 1994-2003.

INTRODUCTION Although county-level obesity estimates are necessary for planning and evaluating community-based interventions, the quality of these data has never been examined. OBJECTIVES To evaluate the reliability of the county-level obesity prevalence estimates from Wisconsin's 72 counties and to highlight the variation of obesity among Wisconsin counties. METHODS Obesity prevalence data for each county in Wisconsin were obtained from the Wisconsin Behavioral Risk Factor Surveys (BRFS) from 1994 to 2003. During this 10-year period, 26,635 residents were interviewed by telephone, with sample sizes ranging from 6586 in Milwaukee County to 15 in Menominee County. The number of counties with reportable and reliable estimates, using criteria of sample sizes > or = 50 and > or = 300, respectively, was determined. RESULTS The 10-year obesity prevalence was reportable for 68 of Wisconsin's 72 counties, ranging from 9.7% in Bayfield County to 29% in Langlade County. By pooling data from the BRFS for 5-, 3-, and 1-year periods, estimates are reportable for 43, 24, and 4 counties, respectively. A sample size of at least 300 provides a more reliable estimate, but is available for only 5 counties for a 5-year period. CONCLUSIONS By pooling 10 years of survey data, obesity rates can be estimated for most of Wisconsin's 72 counties, demonstrating marked variation in rates across the state. This surveillance system provides valuable data for larger counties for planning and program evaluation. Supplemental surveys can be conducted to provide more reliable and timely estimates.