Chronic Conditions – Diabetes Incidence (Adult – Trends)

Data from CDC’s Behavioral Risk Factor Surveillance System (BRFSS) and from the U.S. Census Bureau’s Population Estimates Program were used to obtain county-level estimates of diagnosed diabetes, newly diagnosed diabetes, obesity, and physical inactivity. The BRFSS is an ongoing, monthly, state-based telephone survey of the adult population aged 18 years or older that provides state-specific information on behavioral risk factors and preventive health practices. Respondents were considered to have diagnosed diabetes if they responded “yes” to the question, “Has a doctor ever told you that you have diabetes?” Women who indicated that they only had diabetes during pregnancy were not considered to have diagnosed diabetes. People who reported having diagnosed diabetes were then asked at what age they were diagnosed. Responders were considered to have been diagnosed with diabetes in the last year if they reported having diagnosed diabetes and the difference between their age at the time of the survey and the age they provided to the question, “How old were you when you were told you have diabetes?” was less than one. If the difference was between one year and two years, the person was weighted as half a newly diagnosed case. Respondents were considered obese if their body mass index was 30 or greater. Body mass index (weight [kg]/height [m]2) was derived from self-report of height and weight. Respondents were considered to be physically inactive if they answered “no” to the question, “During the past month, other than your regular job, did you participate in any physical activities or exercises such as running, calisthenics, golf, gardening, or walking for exercise?”

Three years of data were used to improve the precision of the year-specific county-level estimates of diagnosed diabetes and selected risk factors. For example, 2003, 2004, and 2005 were used for the 2004 estimate. Estimates were restricted to adults 20 years of age or older to be consistent with population estimates from the U.S. Census Bureau. The U.S. Census Bureau provides year-specific county population estimates by demographic characteristics—age, sex, race, and Hispanic origin.

County-level estimates for the over 3,100 counties or county equivalents (e.g., parish, borough, municipality) in the 50 US states and the District of Columbia (DC) were based on indirect model-dependent estimates using Bayesian multilevel modeling techniques. This model-dependent approach uses a statistical model that “borrows strength” (a.k.a., Small Area Estimation) in making an estimate for one county from BRFSS data collected in other counties. For incidence rates of newly diagnosed diabetes, multilevel binomial regression models with random effects of demographic variables at the county level were developed. County-level prevalence was based on design-assisted model-based estimates using the power prior log-weights (PLOW) technique developed by Xie et al. Unique PLOW advantages include 1) using single-year BRFSS data rather combining years; 2) inclusion of historical data to define informative priors (power prior); 3) the integration of adjusted sample weights to account for BRFSS’ complex survey design; and 4) more timely estimates with smaller variance. Estima