BRFSS data are often used to rank states or other jurisdictions. Researchers, government officials, and the popular press pay much attention to shifts in these ranks and the highest and lowest ranked jurisdictions. Despite the widespread attention, the rank comparisons rarely incorporate any measure of uncertainty. Standard errors for BRFSS data are available. Using a simple Monte Carlo method, these standard errors can be used to calculate confidence intervals for a rank This approach was used to estimate confidence intervals around a state’s rank for several BRFSS indicators. The resulting confidence intervals were extremely wide, often spanning more than one quartile. For example using this method, the 95% confidence interval for the prevalence of hypertension in Virginia covered 13 ranks. When data are stratified by age, race. or ethnic sub category, confidence intervals are even wider. While rankings are an attractive and popular way of presenting BRFSS data, much caution is needed for their proper interpretation. Close examination of the underlying data, and alternative analyses such as longitudinal analysis or comparisons with standards are probably more useful.
Learning Objectives: Understand the impact of uncertainty on rank orderings. Be able to roughly assess the size of the confidence interval around a rank estimate Describe the limitations associated with the use of rankings and alternative approaches to their use
Keywords: Deaf Patients, Statistics
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.