Department of Epidemiology and Biostatistics

CWRU

Thursday, May 1,1997

3:30 pm - refreshments

4:00 pm - talk

Room 327, Yost Hall

Abstract

Work disability is a serious public health problem. National estimates from the 1990 U.S. census are that 4.2% of those 16-64 years old have a severe work disability, and 8.2% have any work disability. However, work disability data from the census may not give accurate or timely health assessment for policy decisions between census years. While yearly national estimates from surveys, such as the NHIS, are available, they may not apply to local areas. Local health assessment and work disability policy ideally is done using local data. However, most surveillance systems such as NHIS or the BRFSS are done at the national or state level. The usefulness of these data to local health officials may be limited because of the small sample sizes at the local level. To help fill this data void, small area analysis (SM) has been developed, but never applied to estimate local work disability using the BRFSS. Therefore, we studied the usefulness of SM to estimate work disability prevalence rates in the US at the health service area (HSA) and county level. Data included the 1993-95 national BRFSS; the BRFSS was linked to county-level auxiliary data from the 1990 US Census, Area Resource File, and mortality statistics. Definitions of severe work disability were developed from questions in the BRFSS and internally validated. A nested logistic regression model was used to calculate an age-sex-race weighted work disability prevalence rate for each HSA and US county. Validity of model estimates was evaluated directly against BRFSS rates in HSAs and counties with large (>300) BRFSS sample sizes. Mapping the rates revealed the highest prevalence in Appalachia, the deep south, and the northwest. The least work disability was found in the midwest and Rocky Mountains. Validation statistics showed high correlation between the SM and direct BRFSS rates. We conclude that SM is a helpful technique to expand the usefulness of the BRFSS to estimate work disability for local health areas.