Abstract:
|
The Asthma Call-Back Survey (ACBS) is conducted with Behavioral Risk Factor Surveillance System (BRFSS) respondents who report an asthma diagnosis. To evaluate the effects of cell phone samples on prevalence estimates of selected asthma indictors, we reweighted and analyzed five state's 2011-2013 data from the BRFSS combined landline and cell phone survey, ACBS combined landline and cell phone survey, and the ACBS landline only survey respectively. Criteria to assess the effects of adding the cell phone sample included prevalence estimation (PE) for 12 asthma indicators by demographic characteristics, the mean square error (MSE) for bias and variance, and relative standard error (RSE) for reliability. Adding the cell phone samples did not change the PE significantly for most indicators. A lower prevalence of current asthma among children ages 0-4 years with lifetime asthma was seen when results were compared with BRFSS estimates (69% vs.78%). However, adding cell phone samples reduced the bias and variance estimation by reducing the MSE (range: 6% to 48%), and improved reliability by reducing the RSE (range: 7.7% to 22%), depending on the indicators used.
|