A comparison of household and medical provider reported health care utilization and an estimation strategy to correct for response error

Abstract

In health care surveys similar to the MEPS, the use of additional administrative data and medical records for survey participants permits additional methodological investigations and evaluations to examine the accuracy of household reported data. When differentials are observed in the response profiles through these evaluations and comparisons, the design permits well specified adjustment and estimation strategies to correct for measurement error. In addition to serving as the primary source for the expenditures in the MEPS, the design of the Medical Provider Component provides data that could potentially facilitate adjustments to household reported utilization data that correct for reporting errors (both under-reporting and over-reporting(telescoping errors)), under the assumption that the medical provider reports are the gold standard. In this paper, we examine the level of concordance between household and medical provider utilization reports. An adjustment strategy to correct for response error attributable to household utilization reports is also presented.