Statistical Methods in Data Harmonization

• Activity Number: 430
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Quality and Productivity Section
• Author(s): Yan Wang* and Jian Zhang and Honghu Liu
• Companies: UCLA and Kaiser Permanente Southern California and UCLA
• Keywords: Measurement ; Quality ; Harmonization

Abstract

The quality and generalizability of measures is always an issue before any data management steps and statistical models techniques may have taken effect. Data harmonization is referred as a process that attempts to align the data sources that are collected separately but measure the similar contents together, in order to compare and stratify in the analysis. However, this process is very time consuming in practice. For example, the data central wants to compare the performance and cost of different centers all over United states. Data from different centers often have large variability and data collected with various instruments usually have low comparability, even if they are in attempt to measure the same concept or construct. Data harmonization aims to promote common measure for the key indicators that can permit certain degrees of comparability over time and across studies. We will review the statistical methods to harmonize the measures and data. Eventually to improve the quality and comparability of the measurements.