The quality of health data obtained in a new survey of elderly Americans: a validation study of the proposed Medicare Beneficiary Health Status Registry (MBHSR)

Abstract

The Medicare Beneficiary Health Status Registry (MBHSR) is a proposed new survey program that would collect health status indicators annually from large probability samples of Medicine beneficiaries. For reasons of economy, the MBHSR would use mail survey procedures with telephone follow-up of nonrespondents. Because of concerns about response rates and the validity and reliability of the data obtained by such methods, a large-scale (N = 1,922) field test was conducted. The field test assessed the validity of MBHSR survey reports of past medical treatment and conditions by comparing those reports with Medicare claims data. It assessed the (internal) reliability of MBHSR survey responses by comparing responses with logically related survey questions from the MBHSR. Analyses indicate that the MBHSR survey procedures using a combination of mail data collection with telephone follow-up of nonrespondents produced relatively high levels of sensitivity and specificity in identifying medical treatments and procedures previously recorded in Medicare claims data. In addition, the MBHSR Field Test obtained, in general, relatively high levels of internal consistency in survey reports.