Journal of Official Statistics, Vol.12, No.1, 1996. pp. 47–61
Estimates of National Hospital Use from Administrative Data and Personal Interviews
Sally C. Stearns, Mary Grace Kovar, Kevin Hayes, and Gary G. Koch
Abstract:Data sets that merge claims data with health surveys are a rich source of information for estimating hospital use. However, comparisons of claims data with reported use for one survey (the Longitudinal Study of Aging) show that claims data may systematically understate use by Medicare beneficiaries in the United States because of processing lags or programmatic factors. The comparisons also show that self-reports are consistently closer to use measured by claims than are proxy reports, even though survey participants were age 70 or older. Standard techniques are used to adjust national estimates based on the discrepancy between self-reports and claims data, and a composite standard error for the adjusted estimate is derived based on the complex survey design effect and the value of the adjusted estimate. Although the specific correction factor would not apply to other data sets, the process for detection of discrepancies between survey and administrative data and of adjusting for such discrepancies (when resolution of the discrepancies is not possible) has general applicability.
Keywords:Longitudinal Study of Aging; self-reports; proxy reports.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group