Integrating Multiple Sources of Data: Results of Linking Survey and Housing Data

• Activity Number: 445 - GOVT CSpeed 2
• Type: Contributed
• Date/Time: Thursday, August 12, 2021 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Carla E Zelaya* and Lisa B. Mirel
• Companies: CDC/NCHS and CDC/NCHS/DAE
• Keywords: Data Linkage; NHIS; HUD; Administraive Records; Data Integration; Linkage Methodology

Abstract

The National Center for Health Statistics’ Data Linkage Program continually works to enhance their linkage algorithms to minimize type I and type II errors associated with integrating multiple data sources. Recently, an enhanced linkage methodology was implemented for the integration of National Health Interview Survey (NHIS) data and US Department of Housing and Urban Development’s (HUD) administrative records for the largest housing assistance programs, supporting research on the relationship between health and housing. Previously linked NHIS and HUD data relied on deterministic methodologies which resulted in about 9.3% of sample adults and children from the NHIS from 1999-2016 being linked to HUD. The updated linkage relied on both deterministic and probabilistic techniques and resulted in a linked file with a greater number of participants being linked to HUD for overlapping years. This talk will discuss the enhancements to the algorithm, including additional participants being included in the linkage process and the checks that were implemented to assess the quality of the updated linked files. In addition, we will highlight the implications for secondary data analyses.