|
Abstract:
|
The Medicare Current Beneficiary Survey (MCBS) is a continuous longitudinal survey of a nationally representative sample of the Medicare population, including interviews with facility staff on behalf of beneficiaries living in long-term care facilities. Interviewers and facility staff obtain selected survey items from facility and medical records, many of which are redundant to the administrative data certified facilities regularly report to the Centers for Medicare and Medicaid services (CMS), such as the Minimum Data Set (MDS). To reduce burden, the MCBS instrument was redesigned in 2019 to skip these items if a CMS Certification Number (CCN) can be used in record linkage is reported during the interview.
The CCN, present in about half of cases, is used in a deterministic record linkage protocol during data processing and administrative data are substituted for skipped variables to create a blended data product with survey data. This paper evaluates the quality of these blended statistics when compared to past years' data that are abstracted from records or reported in the survey. We assess the accuracy and completeness of the record linkage protocol, data comparability, and comparative changes in levels of item non-response. Overall, this protocol resulted in a near-complete match rate with largely comparable blended data alongside significant operational gains.
|
