Recommendations for Assessing and Evaluating Variable Crosswalks

• Activity Number: 295 - SPEED: Big Data, Small Area Estimation, and Methodological Innovations Under Development, Part 1
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Mitch Sevigny* and Jessica Ketchum and David Mellick
• Companies: Craig Hospital and Craig Hospital and Craig Hospital
• Keywords: Variable Crosswalk; Equating; Linking; Depression

Abstract

Many ongoing long-term research databases are replacing traditional legacy outcome measures with patient reported outcome measures. While these new instruments may do a better job at measuring the construct of interest, data from participants using the legacy measure are lost. Researchers often want to be able to convert data from legacy measures into the new measures using a variable crosswalk. For example, in the Traumatic Brain Injury (TBI) Model Systems National Database, PHQ-9 has historically been used to measure depression. This instrument was recently replaced with a more psychometrically sound instrument, the TBI-QOL Depression Scale. A variable crosswalk was proposed to equate PHQ-9 scores to TBI-QOL scores. A proper crosswalk would certainly be a valuable tool; however, it is essential for researchers to fully evaluate the crosswalk before implementation as they may not be reliable or properly convert scores. This presentation will provide researchers with recommendations for evaluating variable crosswalks for both group level and individual level comparisons. These guidelines will be demonstrated utilizing the proposed crosswalk between PHQ-9 and TBI-QOL.